Developer II

January 2011 Release

QlikView Version 10.00 English

Copyright 2011 QlikTech International AB, Sweden.

Under international copyright laws, neither the documentation nor the software may be cop-ied, photocopied, reproduced, translated or reduced to any electronic medium or machine-readable form, in whole or in part, without the prior written permission of QlikTech Interna-tional AB, except in the manner described in the software agreement.

QlikTech and QlikView are registered trademarks of QlikTech International AB.

Microsoft, MS-DOS, Windows, Windows NT, Windows 2000, Windows Server 2003, Win-dows Server 2008, Windows XP, Windows Vista, SQL Server, Excel, Access, Visual Basic, Internet Explorer, Internet Information Server, Visual C++, Visual Studio and MS Query are trademarks of Microsoft Corporation.

IBM, AS/400 and PowerPC are trademarks of International Business Machines Corporation.

Firefox is a trademark of the Mozilla Foundation.

Apple, iPhone, iPod Touch, Safari and MacOS is a trademark of Apple Corporation.

BlackBerry is a trademark of Research In Motion.

This release: January 2011

CONTENT1 INTRODUCTION 7

Installing the Course Materials 7Program versions 7Text formats 7

2 DEVELOPER TOOLS 9

Troubleshooting 9The Debugger 9

3 MAPPING TABLES 19

Mapping Quarters to the Orders table 19MonthYear 23Cleaning up the table structure 26

4 LOADING BUDGET DATA 31

Reading Cross Tables 31

5 ADVANCED SCRIPTING 39

Condition on a field in a table 39Aggregating Data 39Joining tables 39Concatenation 40Preceding Load on Preceding Load 41

6 ADVANCED CALCULATIONS IN SHEET OBJECTS 51Set Analysis 51Dollar-Sign Expansion 54AGGR Function 56

QlikView Developer II | CONTENT7 SCRIPTING AND DATA MODELING CHALLENGES 63

Link Tables and Concatenated Tables 63Calculating net change within a field 63Functions InDate, Data & DateIslands 64Aggr() 64Class() 65Dynamic Aggregation 65IntervalMatch 65

8 DATA MODEL OPTIMIZATION 81

Performance Tuning 81Best Practices for QlikView File Optimization 83Creating Incremental Loads 85

9 DAILY AND TRANSACTION BALANCES 95

Overview 95Example: Balances at Specific Dates 95Example: Balances for Transactions (in/out) 98

10QLIKVIEW SECURITY 101

Access control 101Access levels 101Access control database 102Inherited access restrictions 103Hidden script 104Adding Section Access 104Access control for certain commands 109Further access control 110Unattended Command Line Reload Considerations 111Access restrictions on selected field values 111Field value limitation in Section Access 112

11ADVANCED DATABASE CONNECTIVITY 115

Custom Data 1154

SAP Connector 116

QlikView Developer II | CONTENT12BUSINESS CASE WORKSHOP 119

Scenario 119Data Structure 1215

QlikView Developer II | CONTENT6

QlikView Developer II | 1 INTRODUCTION1 INTRODUCTIONMany students will come to this course having already taken QlikView classroom training. This chapter covers formatting and coventions used in the course manual.

1.1 Installing the Course MaterialsThe course materials will self-extract from the file into the default directory

C:\QlikViewTraining\DeveloperII\

Make a Windows shortcut to this folder and place it on your desktop.

Also make a Windows shortcut to the documentation folder and place it on your desktop.

C:\Program Files\QlikView\DocumentationorC:\Prrogram Files (x86)\QlikView\Documentation

1.2 Program versionsThis course was created using the English version of QlikView 10.00 run-ning on Windows7. If other operating systems or languages are used, minor differences may be noted in the visual appearance of windows and dialog boxes.

1.3 Text formatsExercises and actions to be completed by you, the student, will be set-off with a logo, as you see, below:

Exercise/Do:This is a sample of instructions you would see to complete an exercise con-taining a sequence of steps.

1 Click on the Start button2 Locate the QlikView icon3 Click on the QlikView icon to launch the program

All commands, as well as all names of menus, dialogs and buttons are in the following font style: File - Open

All names of list boxes, graphs and specific data in list boxes, etc. are in the following font style: Country

All file names are in the following font style: QlikViewCourse.qvw7

QlikView Developer II | 1 INTRODUCTIONTips and Notes are outlined in a highlighted box, as you see below:

This sample sentence is used to illustrate important points in the text, tips and notes to consider as you complete the course materials8

QlikView Developer II | 2 DEVELOPER TOOLS2 DEVELOPER TOOLS

2.1 TroubleshootingDebugging is an art. There are best practices and guidelines that can help in the debugging process. The key point to take away from debugging is to change one item at a time and then examine the impact of that change. If more than one item is changed without an examination, debugging becomes difficult. This portion of the class will not teach you how to debug an docu-ment but will point you in the direction of where to look if your document does not operate or look like you expected it to. These basic tools and pro-cesses are the best place to start when debugging a QlikView document.

Also, during script execution in debug mode, the debugger allows for moni-toring of variables. This is useful for validation of variable state at different stages of script execution.

Remember, always make sure the Generate Logfile checkbox is checked under Settings|Document Properties|General. Also remember, missing or misplaced commas and semicolons are the source of many error messages. Check for these common problems first.

2.2 The Debugger

Where is that Breakpoint, Check the Variables

Breakpoints stop script processing to give you a chance to troubleshoot. They can be placed in your script with the Debug option in the QlikView Script Editor. Breakpoints only stop script processing if placed on the line of the first statement in a code block that is terminated by a semi-colon. Setting a breakpoint on any other statement in the code block has no effect on breaking the script as it runs.

Objectives Understand the importance of troubleshooting Use the debugger Introduce additional tools and methods9

QlikView Developer II | 2 DEVELOPER TOOLSYou will notice in the Figure above that three breakpoints are set at lines 38, 39 and 44. Note that the debugger will not stop at line 39 since it is not the first statement in the code block.

This could be a LOAD statement, a SQL LOAD with a preceding LOAD, or a grouped LOAD statement. If you trace each statement back from the semi-colon you will see that while the key word MAPPING on line 39 is part of a code block ending in a semi-colon, MAPPING is not the first line of the code block Shippers: is.

2.2.1 Variables

If you have created variables in the load script, you can view the value of the variable as the script is loading. You might be surprised that it does not hold the expected value(s). Create additional variables to investigate more states during script loading.

Figure 1. Breakpoints10

QlikView Developer II | EXERCISEExercise: Debugging

Do: 1 Navigate to the C:\QlikViewTraining\DeveloperII\Chapter02 folder.2 Load the file Debugging.qvw into Qlikview and try loading data into

QlikView after setting multiple breakpoints, step through the code and see where the debugger stops. Also, check the values of the variables

3 Open the file \Debugging\Variable.qvw. Enter the debugger from the script editor and check the breakpoints that have been set in the script

4 Use the Run button to execute the script up until the following break-point.

5 Note that the variable mySetVar is not listed in the bottom right variable monitor pane.

6 Hit Run again and note that the value of the mySetVar variable has been set.

7 Hit Run once more and exit the script.8 Repeat steps 3 to 6 and note on step 4 that the variable mySetVar still

retains the value from the previous execution.9 Go to Settings | Variable Overview and note that the script variable

mySetVar (once created) is stored and accessible from within QlikViewDesktop.

Remember, intermittent values of script variables can be monitored through the debugger. The scope of script variables is not limited to a single script execution, which is to say the variables persist between executions.11

QlikView Developer II | 2 DEVELOPER TOOLS2.2.2 Using the Table Box Sheet Object

The Table Box is the developers best friend. It is great for identifying how data relates and where data might be missing.

The table object is great for viewing row-level data relationships between different fields. For ease of use, a table box may have dropdown select enabled from the Presentation tab enabling the developer to quickly navi-gate each field / column and make text query selections.

Creating a Table Box is easy and one can add columns from multiple tables as well as add dropdown selects to the columns (fields) in the table. Note the Dropdown Select checkbox in the Figure below; the dropdown select allows you to select and search for data in the Table Box. This will help you under-stand how the data relates across your tables.

You can also see where NULL values happen across tables and within fields.

NULL values can adversely affect a document and the selection of data.

Figure 2. Adding Dropdown Select12

QlikView Developer II | EXERCISEExercise

Do:1 Navigate to the C:\QlikViewTraining\DeveloperII\Chapter02 folder.2 Open the document KeyFields.qvw and go to the Main tab where you

will see a table. Select a cell that holds data and note that any row with an identical field value will be displayed with the corresponding columns of data with which it is associated.

3 Next, clear the selections (the Clear button) and select a cell that has a null (the default representation is a character). Note that nothing happens.

Note: You are seeing disjointed information, which can be problematic. You can explore the causes of this condition by examining key fields.

Remember, Table boxes are great for understanding and validating data relationships. Only included / possible values are displayed in a table box.13

QlikView Developer II | 2 DEVELOPER TOOLS2.2.3 Making Copies of Key Fields

How might this disjointed condition have been created? A LEFT JOIN A MAPPING table operation Bad data Key field mismatches

Key Fields connect tables in QlikView, but examining them will not always display the information you might expect. Showing a Key Field in a Sheet Object such as a Table Box, displays all possible values from the Key Field in any and all connected tables any table(s) could contain distinct values that do not match any values in any other table(s).

To get an understanding of what is happening in every table on every side of a Key Field, load a copy of the Key Field in every table and uniquely rename each Key Field copy with an AS clause.

Reload your document, create List Boxes for all fields, then do a SELECT ALL on one of the fields that is an alias of the Key Field.

Finally, SELECT EXCLUDED on the Key Field by right clicking the Key Field List Box. You can see, through the selection and exclusion, which items are joined through the Key Field and which are not.

This problem usually happens when there is missing data. You can also return erroneous results if you are looking at different discreet levels of information in the model. This condition is covered later in the course.14

QlikView Developer II | EXERCISEExercise

Do:1 Navigate to the C:\QlikViewTraining\DeveloperII\Chapter02 folder.2 Open the document KeyFields.qvw3 Comment the script on the Standard Join tab and uncomment the script

on the Copy of Keys tab.4 Save and reload.5 Note that the key field in each table has been renamed uniquely, i.e., Key

AS Key1_Table1.6 Create List Boxes for all fields.7 SELECT ALL on one of the aliased key fields.8 SELECT EXCLUDED on the Key field (right click in the Key field List

Box and choose SELECT EXCLUDED).9 Observe the matches and mismatches between the keys in each table.

Debugging: Examples and Steps

When building and debugging complex equations, take two approaches:

1 Take a sample of the data and do not create a data model that is more that two tables, if possible. There should not be more than three col-umns in each table or more that fifteen rows. Use the inline wizard or Excel to create the tables.

2 When building complex expressions, use a Straight Table chart and build expressions a single column at a time.

Do:1 Navigate to the C:\QlikViewTraining\DeveloperII\Chapter02 folder.2 Open the ComplexEquations.QVW. 3 Create a Straight Table chart with CompanyName as the dimension plus

the following expressions:Expression Label Sum(LineSalesAmount) SalesCount(LineNo) LinesAvg(LineSalesAmount) AvgSalesSum(Quantity) Qty

Note: By the use of the 'Select all', 'Select possible' and 'Select excluded' selection modes on key copy fields, it is possible to investigate joined and disjointed data within your data structure.15

4 Finish with 2 more expressions:

QlikView Developer II | EXERCISEExpression Label Qty/Lines AvgQtySales/Qty UnitPrice

Here you see that you can type an expression directly or you can use an expression label as an alias to the expression to get the same results.

5 Build your complex expression one column at a time and then move the expression to the placeholder of the Label that is referenced in the expression. This is the best approach for slowly building complex expressions.

Dates

In many cases, date fields are key fields. To join tables from different sys-tems, there are cases where dates need to be converted to a common format before dates from disparate datasources will match properly. It is always a good idea to check the key field using a Table Box to verify that the date for-mat is consistent throughout the entire key field.

Do:1 Navigate to the C:\QlikViewTraining\DeveloperII\Chapter02 folder.2 Open a working copy of the Dates.qvw file and load the data from the

Dates.XLS file in the Datasources folder. Create two tables in the script. Your script should look like this:

Note: You can also reference an expression by its Table Column number (zero-based index) but DONT DO IT!

Note: You cannot nest aggregations (aggregate an aggregation) except by using the aggr() function, covered later in this course.16

QlikView Developer II | EXERCISE3 Next, create a Table Box including all columns, sort by Date ascending, and scan the data. You should see something like the Table Box below. Note that the dates look the same but similar dates do not link.

4 To determine what is causing the problem, create a Text Object that dis-plays the expression

=num(Date, ###.######)and alternately select each of the two similar dates. Note the fractional part of each displayed date in the Text Object. The date from one table is actually a timestamp while the other date has no time associated with it.17

QlikView Developer II | EXERCISE5 Next fix Date2 with the floor() function (floor(Date2)) changing Date2 from a timestamp to a date. (Note the floor() function truncates a deci-mal number and returns the next lower integer value, i.e., 2.7 becomes 2). Now reload the data. The result should look like this where you have rows that now link.

6 As an alternative, try using the daystart() function on the dates in both tables. Verify the results are the same. Hide the fact that both dates are now timestamps starting at midnight.

Note: Be wary when comparing dates. Ensure that the granularity is the same and stored in the same manner in each date field. (e.g. Day level dates stored without a time component, month level dates stored as the first day of each month, etc.18

QlikView Developer II | 3 MAPPING TABLES3 MAPPING TABLES

Sometimes you need to add an extra field to a table to use a combination of fields from different tables, or you want to add a field to clean up the data structure. QlikView has an effective way to add single fields to a table called mapping tables. In this chapter, we will take a look at how mapping tables work.

3.1 Mapping Quarters to the Orders tableThe Quarters table is useful, in that it links our Month data in the Orders table with the correct Quarter. However, the Month field is now a key field, and this will probably cause problems later. The following illustrations give us a visual of this dilemma:

Objectives Understand mapping tables Use mapping tables to add Quarters to the Orders table Clean up the table structure

Figure 1. Quarters Table key link on Month.19

QlikView Developer II | 3 MAPPING TABLESBy changing our Quarters table into a MAPPING table, we will be able to integrate the Quarters field into the same table as Month (the Orders table).

The MAPPING prefix is used on a LOAD or SELECT statement to create a mapping table. Tables read via MAPPING LOAD or MAPPING SELECT are treated differently from other tables. They will be stored in a separate area of memory and used only as mapping tables during script execution. After script execution they will be automatically dropped.

A mapping table must have two fields, the first one containing comparison values and the second the desired mapping values. The two fields must be named, but the names have no relevance in themselves. The field names have no connection to field names in regular input tables. When mapping tables are used to map a certain field value or expression, that value will be com-pared to the values in the first field of the mapping table. If found, the origi-nal value will be replaced by the corresponding value in the second field of the mapping table. If not found, no replacement is made.

The syntax is:mapping ( load statement | select statement )

Figure 2. Month field with key indicator in the Available Fields listing.20

QlikView Developer II | EXERCISEExercise

Do:1 Launch QlikView and save a working copy of the MappingTables.qvw

QlikView file in the working directory for this chapter (C:\QlikViewTraining\DeveloperII\Chapter03).

2 Open the Edit Script dialog.3 Now, let us change the Quarters table load into a mapping load. 4 On the Main tab uncomment the Quarters table load statement you

want to use. Make sure that the other table is still commented so that you do not read from two Quarters tables.

5 Add _Map to the table name.6 On the next line, type MAPPING in front of the LOAD statement.7 When complete, verify that this section of your script resembles the fol-

lowing:

Quarters_Map:MAPPING LOAD

rowno() as Month,'Q' & Ceil(rowno()/3) as Quarter

Autogenerate(12); Do not save and close just yet. If you reload the data now, you will lose the Quarters table and field, since mapping tables are temporary. How-ever, we can use the Quarters_Map table in our script (as long as we use it after it is defined in the script). To do this, we will use the applymap function.

The syntax is: applymap( 'mapname', expr, [ , defaultexpr ] )

The applymap function maps any expression on a previously loaded mapping table. Mapname is the name of a mapping table previously loaded by a MAPPING LOAD or MAPPING SELECT statement. The name must be quoted with single quotes. Expr is the expression whose result will be mapped.Defaultexpr is an optional expression, which will be used as the default mapping value if the mapping table does not con-tain any matching value for expr. If no default is provided, the value of expr is returned is unchanged.

8 Add an applymap function to the Orders table, based on the numeric value of Month. This function should refer to the Quarters_Map table. Refer to the syntax example that follows:

applymap('Quarters_Map',num(month(OrderDate)),null()) AS Quarter,21

QlikView Developer II | EXERCISE9 Save and Reload the document. 10 Open the Table Viewer to verify the Quarters table is gone and that

there is now a field called Quarter in the Orders table.

Figure 3. Using ApplyMap to embed fields into another table22

QlikView Developer II | 3 MAPPING TABLES3.2 MonthYearWe will complete our time dimension fields by creating a new field that makes every month unique. There are, of course, several ways to accomplish this. In this course, we will create the field MonthYear by using QlikView date functions based on the OrderDate field, along with a date formatting function to provide the correct display format for our new month field.23

QlikView Developer II | EXERCISEExercise

Do:1 Return to the MappingTables.qvw file you have been working on in this

chapter.2 Open the Edit Script dialog from the menu or toolbar.3 Locate the Orders table LOAD statement.4 Immediately following the applymap as Quarter field line, create a new

field named MonthYear in the LOAD statement for the table Orders, as follows:

date(monthstart(OrderDate),'MMM-YYYY') AS MonthYear,The monthstart function returns the first day of the month of the Order-Date value. The date function then formats this value into a 3-character month name, followed by a 4-digit year. Since QlikView stores this field as both a text string (the format we just specified) and as a numeric, it can be sorted numerically, as you would expect.

The complete Orders table LOAD statement should now look as follows. Be sure your script syntax matches this. Note, your Quarter and Month-Year fields line will likely fit on a single line instead of wrapping as seen below.

//*************** Orders table ***************Orders:LOAD CustomerID,

EmployeeID,EmployeeID AS EmployeeSalesID,Freight,OrderDate,year(OrderDate) AS Year,month(OrderDate) AS Month,day(OrderDate) AS Day,applymap('Quarters_Map',

num(month(OrderDate)),null()) AS Quarter,date(monthstart(OrderDate), 'MMM-

YYYY') AS MonthYear,OrderID,OrderID AS OrderIDCounter, ShipperID;

SQL SELECT * FROM Orders; 24

QlikView Developer II | EXERCISE5 Save and Reload the script.6 Add a List Box for the new MonthYear field to your sheet. 25

QlikView Developer II | 3 MAPPING TABLES3.3 Cleaning up the table structureYou will most likely want to minimize the number of tables in QlikView. (Omit needless tables.) It takes computational resources to make calcula-tions between tables. If you have tables with only two fields, map those tables to another table and so minimize the number of tables. Let us look at the Table Viewer and see if there are any tables than can easily be mapped to another table.26

QlikView Developer II | EXERCISEExercise

Do:1 Return to the file you have been working on in this chapter.2 Open up the Table Viewer.

As we can see in the Table Viewer, some tables have only two fields. These tables could be mapped to the connecting tables. Let us start by mapping the Shippers table to the Orders table.

3 Open up the Script Editor and scroll down.4 Change the table referring to the script for the Shippers table as follows.

Shippers_Map:MAPPING LOAD

ShipperID,CompanyName AS Shipper;

SQL SELECT * FROM Shippers;

5 Add the following line to the bottom of the Orders table.

applymap('Shippers_Map', ShipperID, 'MISSING') AS Shipper

The above line of script tells QlikView to use the word MISSING in the Shipper field where no matching ShipperID values can be found.

Figure 4. The Table Viewer27

QlikView Developer II | EXERCISE6 Verify that your Orders table script should resemble the following;

//*************** Orders table ***************

Orders:LOAD CustomerID,

EmployeeID,EmployeeID AS EmployeeSalesID,Freight,OrderDate,year(OrderDate) AS Year,month(OrderDate) AS Month,day(OrderDate) AS Day,applymap('Quarters_Map',

num(month(OrderDate)),null()) AS Quarter,

date(monthstart(OrderDate), 'MMM-YYYY') AS

MonthYear,OrderID,OrderID AS OrderIDCounter, ShipperID,applymap('Shippers_Map', ShipperID,

'MISSING')AS Shipper;

SQL SELECT * FROM Orders;

7 Save the document and Reload the script.8 Take a look in the Table Viewer to see that in fact the Shipper field is

now a part of the Orders table.

Figure 5. The Shipper field is now part of the Orders table.28

QlikView Developer II | EXERCISEExtra Credit Exercises

Do:1 Launch QlikView and save a working copy of the AdditionalMap-

pingExercises.qvwQlikView file in the working directory for this chapter (C:\QlikViewTraining\DeveloperII\Chapter03).

2 Open the Edit Script dialog.3 Create a tab after the Main tab and name it Mapping Loads.4 Move the Shippers_Map and Quarters_Map script to the Mapping

Loads tab.5 Map the Divisions table to the Customers table. Make sure to remove

(or comment out) the Divisions table from the Dimensions tab and cre-ate a mapping table on the Mapping Loads tab.

Although script examples are at the bottom of this page, they are for ref-erence only should you need help. We encourage you to try to add the Mapping Load and ApplyMap script on your own.

6 Are there any other tables that can be mapped to another table? Check the Table Viewer. Make sure to look for tables with only two fields. Dis-cuss with the course Instructor.

//*************** Divisions ***************Divisions_Map:MAPPING LOAD

DivisionID,DivisionName;

SQL SELECT *FROM Divisions;//*************** Customers ***************Customers:LOAD Address,

City,CompanyName,ContactName,Country,CustomerID,DivisionID,applymap ('Divisions_Map', DivisionID)

as

Division,Fax,Phone,PostalCode,29

QlikView Developer II | EXERCISEStateProvince;SQL SELECT *FROM Customers;30

QlikView Developer II | 4 LOADING BUDGET DATA4 LOADING BUDGET DATA

In our example data, there is a Budget table for Employees and Offices. We are going to load this into our QlikView document. The Budget table is built as a cross table and we need to convert this when we read it into QlikView. We will also add a field to the Budget table that allows us to change the val-ues of the budget to help in planning.

4.1 Reading Cross TablesFirst, we will open the Budget table that is contained in the Excel file. This table does need some rework to read it into QlikView. Fortunately, QlikView has excellent functionality to interpret and change tables so that we do not need to alter the original look of the Excel file. We will use the crosstable wizard to load the data initially, and then make adjustments to our script. We will also use an input field to load the data initially, and then make adjustments to our script.

Objectives Understand cross tables Use the File Wizard to transform the data and create load script31

QlikView Developer II | EXERCISELoading Budget Data Exercise

Do:1 Launch QlikView and save a working copy of the LoadingBudget-

Data.qvw QlikView file in the working directory for this chapter (C:\QlikViewTraining\DeveloperII\Chapter04).

2 Open the Edit Script dialog.3 Create a new tab following the Sales Person tab and call it Budget.4 Click on the Table Files button and open the Budget.xls file from the

data folder for this chapter.5 In the File Wizard, start by setting the Header Size to one line.6 Next, we need to make sure that there are no empty rows in the Office

field. Click on the Next button and then click on Enable Transforma-tion Step to transform the table and then click the Fill tab.

7 Click the Fill button and then Cell Condition. We want the cell to fill if it is empty. Click OK, OK and Next to return to the File Wizard.

8 Click on Crosstable to change the table from a cross table to a nor-mal table.

9 Click on the Qualifier Fields and type 2 to expand them to include both the Office and EmployeeID fields in purple.

A qualifying field in a cross table, is a field that should not be altered during the Cross table load.

2006 is not a qualifying field. This is the first of the fields we want to transform so that the years are placed in one field and the budget values are placed in another field.

10 Name the Attribute BudgetYear.11 Name the data BudgetAmount.12 Click OK.33

QlikView Developer II | EXERCISE13 Click Finish. You should have the following table in the script.

CROSSTABLE(BudgetYear, BudgetAmount, 2)LOAD

Office,EmployeeID,[2006],[2007],[2008],[2009],[2010]

FROM Datasources\Budget.xls (biff, header is line, embedded labels, table is [Sheet1$], filters(Replace(1, top, StrCnd(null))));

14 Name this table BudgetTemp.15 Save and Reload the document.

Open the Table Viewer. As you can see, there is a synthetic key between the BudgetTemp and the Employees table. We want to remove this syn-thetic key. We will adjust our script to get the data we need from the BudgetTemp table, and then we will drop it.

16 Go to the Script Editor.34

QlikView Developer II | EXERCISE17 For simplicitys sake, change the BudgetTemp table to the following script by replacing the individual field names with the asterisk. This will allow your script to run properly even when the labels in the data change.

BudgetTemp:CrossTable(BudgetYear, BudgetAmount, 2)LOAD

*

FROM

[Datasources\Budget.xls] (biff, embedded labels, header is 1 lines, table is Sheet1$, filters(Replace(1, top, StrCnd(null))));

18 To fix the first part of the synthetic key issue, go to the File Data tab and add the following line to the top of the Employees table.

Office&'-'&EmpID AS BudgetKey,19 Save and Reload the script.

Do:1 Now we will address the synthetic key field in the Budget statement and

add the input field to set the budget prognosis. If it is not already open, launch the QlikView document you have been working on in this chap-ter.

2 Go to the Script Editor and place the cursor right after the SET state-ments on the Main tab.

3 Enter the following statement.

INPUTFIELD BudgetPrognosis;

The INPUTFIELD statement tells QlikView that the field will be an INPUT field. You have to state this in the script before you actually read the field in a table.35

QlikView Developer II | EXERCISE4 Go to the Budget tab and create a Budget table, as follows:

Budget:LOAD

Office &'-'& EmployeeID AS BudgetKey,BudgetYear,BudgetAmount AS BudgetPrognosis,BudgetAmount

RESIDENT BudgetTemp;

DROP TABLE BudgetTemp; 5 Save and Reload the script.

Now, we can use the INPUT field BudgetPrognosis to set different bud-get values if we need to alter the budget to correspond to the actual val-ues.

6 Create a new Table Box titled Sales Budget consisting of the following fields: SalesPerson, BudgetYear, BudgetAmount, and BudgetPrognosis.

7 Move your mouse cursor over the BudgetPrognosis column in the table box. An entry arrow icon will appear.

8 Click on the Input icon on any row and enter any number.9 Right-click on the BudgetPrognosis column. Notice the related options

of Restore Single Value, Restore Possible Values, and Restore All Values.36

QlikView Developer II | Now, we can use the INPUT field BudgetPrognosis to set different bud-get values if we need to alter the budget to correspond to the actual val-ues.37

QlikView Developer II | 5 ADVANCED SCRIPTING5 ADVANCED SCRIPTINGThere are several key measures to create in this chapter. We need to calculate these in the script. They are OrderLineAmount, CostOfGoodsSold and Margin. To make these calculation fields, we need to do some advanced scripting. There is also a key field, CategoryType, which we need to create.

5.1 Condition on a field in a tableThe CategoryType field can be created by using the CategoryID field. If the CategoryID is 5 or 6, the CategoryType should be Footwear, otherwise the type should be clothing. Let us create this field in the script.

5.2 Aggregating DataOne of the key measures for this chapter is the OrderSalesAmount. We need to calculate this in the script. At the moment we already have the LineSale-sAmount, but we want to have a total amount for each Order. To accom-plish this, we need to aggregate the LineSalesAmount.

To group or aggregate data, we will use the GROUP BY clause in the LOAD statement. In this case, we need to aggregate the data in the OrderDetails table by OrderID.

5.3 Joining tablesWe want to add the OrderSalesAmount field to the Orders table. To do so we can add the values of this table to the Orders table. To use two tables together like this, we must begin by combining them into a single table. Here, the JOIN between tables can be performed against the source data-base or we can use a QlikView JOIN command. Since we already have the source data we need loaded into memory, we will use the QlikView JOIN LOAD statement against the table just created.

ObjectivesLearn and use Conditions in tables Aggregation Joining tables Preceding Load on Preceding Load

Tip: Bad joins can result in significant memory penalties. If you receive an Out of Memory error, be sure to double-check your script and data model!39

QlikView Developer II | 5 ADVANCED SCRIPTING5.4 ConcatenationAnother way to join data together from multiple tables is to use concatena-tion. There are two ways to concatenate data. We will explore each of these methods.

5.4.1 Automatic ConcatenationIf the field names and the number of fields of two or more loaded tables are exactly the same, QlikView will automatically concatenate the results of the different LOAD or SELECT statements into one table.

Example:

LOAD a, b, c FROM Table1.csv;LOAD a, c, b FROM Table2.csv;The resulting logical table has the fields a, b and c. The number of records is the sum of the numbers of records in table 1 and table 2.Rules:

- The number and names of the fields must be exactly the same.

- The order of the fields listed in each statement is arbitrary.

- The order of the two statements is arbitrary.

5.4.2 Forced ConcatenationIf two or more tables do not have exactly the same set of fields, it is still pos-sible to force QlikView to concatenate the two tables. This is done with the Concatenate prefix in the script, which concatenates a table with another named table or with the last previously created logical table.

Example:

LOAD a, b, c FROM Table1.csv;Concatenate LOAD a, c FROM Table2.csv;

The resulting logical table has the fields a, b and c. The number of records in the resulting table is the sum of the numbers of records in table 1 and table 2.

The value of field b in the records coming from table 2 is NULL.

Rules:

- The names of the fields must be exactly the same.

- The order of the fields listed in each statement is arbitrary.

- Unless a table name of a previously loaded table is specified in the con-catenate statement the concatenate prefix uses the last previously created logical table. The order of the two statements is thus not arbitrary.40

QlikView Developer II | 5 ADVANCED SCRIPTING5.4.3 Prevent ConcatenationIf two tables have the same set of fields and thus would normally be auto-matically concatenated, you can prevent the concatenation with the NoCon-catenate prefix. This statement prevents concatenation with any existing logical table with the same set of fields.

The syntax is:

NoConcatenate ( LoadStatement | SelectStatement )Example:

LOAD a, b FROM Table1.csv;Noconcatenate LOAD a, b FROM Table2.csv;

In our data, we have been provided with an additional set of new employees that are not yet contained in the EmpOff.xls file. In order to add this data, we need to modify our load script.

5.5 Preceding Load on Preceding LoadThe next key measure we are going to add is the CostOfGoodsSold. To cal-culate this value, we need to add the UnitCost field from the Products table to the OrderDetails table. We are going to do this by using a mapping table and apply this to the OrderDetails table.41

QlikView Developer II | EXERCISEExercises

Do: Condition on a Field in a Table1 Launch QlikView and save a working copy of the AdvancedScript-

ing.qvw QlikView file in the working directory for this chapter (C:\QlikViewTraining\DeveloperII\Chapter05).

2 Open the Edit Script dialog and go to the Dimensions tab.3 Find the Categories table and place the cursor after the last field of this

table. 4 Type a comma and press ENTER to get to a new row. Type the following

to create the CategoryType.

IF(CategoryID = 5 OR CategoryID = 6, 'Footwear', 'Clothing') AS CategoryType;The IF statement in QlikView uses the following syntax:

if( condition , then , else )The condition should be evaluated to be either true or false. If the condi-tion is true, the then part will be processed. However, if the condition is false, the else portion of the statement will be processed.

5 Save the script and Reload. 6 Look at the fields. You can now see that we have a new field named Cat-

egoryType.

Do: Aggregation1 Open the QlikView file you have been working on in this chapter.2 Open the Script Editor and place the cursor after the OrderDetails table

on the Orders tab.3 Add the following statement to your script:

LOAD OrderID,sum(LineSalesAmount) AS

OrderSalesAmount RESIDENT OrderDetailsGROUP BY OrderID;

4 Notice the aggregation function sum(LineSalesAmount) included in this statement. This function will be evaluated over all the potential combi-nations of the other fields in the LOAD (OrderID) statement. The GROUP BY clause is needed to aggregate, or group fields other than those included in the aggregation. In this case, it will total the Sales Amount for each OrderID.43

QlikView Developer II | EXERCISEDo: Joining Tables1 Go to the Script again and place the cursor just in front of LOAD in the

table just created.2 Type LEFT JOIN (Orders) in front of the LOAD statement. The result

should be as below.

LEFT JOIN (Orders)LOAD

Here we use a LEFT JOIN load because we want to make sure that we do not get any values of Orders that do not exist in the Orders table. In QlikView, the default join behavior is a full outer join. Therefore, if there are no matching fields between the two joined tables, you will get a Car-tesian product of the records. Since we are specifying OrderID in both tables, and we are specifying Left, only the records matching OrderID included in the Orders table will be included. We include the OrderSale-sAmount field because that is what we want to add to the Orders table.

3 Save and Reload the script.

Do: Concatenation1 Open the Edit Script dialog in the QlikView file you have been working

on in this chapter.2 Position your cursor on the File Data tab directly after the Employees

table has been loaded. We need to duplicate the fields we currently have for Employees, so we will not use the File Wizard in this case. Instead, copy the Employee LOAD statement, and paste the copied text after the original text.

3 Since the new file data format matches our first file, we only need to change the source of the data. Revise the From clause in the new load statement to read as follows:

FROM Datasources\Employees_New.xls (biff, embedded labels, table is [Employee$]);

4 Click OK and Save the QlikView document.5 Run the script.

If you notice a number of new Synthetic Keys, or a new $Table value of Employee-1, you know something did not work correctly with auto-matic concatenation.You can avoid a number of potential problems with automatic concate-nation by using the concatenate prefix on load statements that you know should be concatenated.

6 Add the concatenate prefix to the new Employee LOAD statement, and specify the Employees table. 44

QlikView Developer II | EXERCISEThis will always concatenate these two tables together, even if inadver-tent script changes are made later to one of the loads, but not the other. The new Employee LOAD statement should now begin as follows:

Concatenate (Employees) Load7 You may have noticed that there are very few differences between our

two Employee LOAD statements. In fact, we can use another QlikView feature to load the same data in just a single load statement. By using a wildcard specification on the FROM file name, QlikView will automati-cally load from all files matching that specification, and concatenate the data into a single logical table for you. Since both our file names start with Emp, and have the .xls file extension, we can use the wildcard Emp*.xls in the FROM clause. If we make this change, and comment the second Employee LOAD statement, the script should now read as fol-lows:

Employees:Load Office & - & EmpID as BudgetKey,

EmpID AS EmployeeID,//[Last Name],//[First Name],[First Name] & ' ' & [Last Name] AS Name,Title,[Hire Date],Year([Hire Date]) AS [HireYear],Office,Extension,[Reports To],[Year Salary]

FROM Datasources\Emp*.xls (biff, embedded labels, table is [Employee$]);//Employees://Concatenate (Employee) Load

//Office & - & EmpID as BudgetKey,//EmpID AS EmployeeID,//[Last Name],//[First Name],//[First Name] & ' ' & [Last Name] AS

Name,//Title,//[Hire Date],//Year([Hire Date]) AS [HireYear],//Office,45

QlikView Developer II | EXERCISE//Extension,//[Reports To],//[Year Salary]

//FROM Datasources\Employees_New.xls (biff, embedded labels, table is [Employee$]);

8 Save the revised script and the QlikView document. Then Reload, and verify the Employee data has not changed.

9 As an optional exercise, you may want to try to determine why the employees listed in the Employees_New.xls file are not assigned e-mail addresses (field e-mail is null for these employees). What do you need to do to correct this problem?

Do: Preceding Load on Preceding Load1 In the QlikView file you have been working on in this chapter, go to the

Script Editor and place the cursor at the bottom of the Mapping Loads tab.

2 Create the following table either by typing it from scratch or by using the Select button.

UnitCost_Map: MAPPING LOAD

ProductID,UnitCost;

SQL SELECT * FROM Products;

3 Go to the Orders tab and add the following script line to the bottom of the OrderDetails table just above the SQL SELECT * line. Remember to remove the semi-colon from the line above and replace it with a comma.

applymap('UnitCost_Map', ProductID, 0) * Quantity AS

CostOfGoodsSold;

We combine the applymap function with a calculation and create the CostOfGoodsSold field directly in the preceding LOAD of the OrderDe-tails table.

The last of the remaining key measures that we need to create in the script is the Margin. The Margin is calculated as the LineSalesAmount CostOfGoodsSold. The easiest way to do this is to place a preceding load on top of the preceding load in the OrderDetails table. You can add several preceding loads on top of each other and they will be evaluated from the bottom and up. This means that you can use a field created in a preceding load in a new preceding load on top of the first one.46

QlikView Developer II | EXERCISE4 We will use this functionality to create the Margin field.5 Put the cursor after the OrderDetails label.6 Create a preceding load by adding the following script.

LOADLineSalesAmount - CostOfGoodsSold AS Margin,*

;7 The full OrderDetails script should look like this:

//*************** Order Details table OrderDetails:LOAD

LineSalesAmount - CostOfGoodsSold AS Margin,*

;LOAD

Discount,LineNo,OrderID,ProductID,Quantity,UnitPrice,UnitPrice * Quantity * (1 - Discount) ASLineSalesAmount,applymap('UnitCost_Map', ProductID, 0) *

Quantity AS CostOfGoodsSold;SQL SELECT * FROM `Order Details`;

LEFT JOIN (Orders) LOAD OrderID,sum(LineSalesAmount) AS OrderSalesAmount

RESIDENT OrderDetailsGROUP BY OrderID;

8 Save and Reload the script.

The new key measure fields should be ready for use.47

QlikView Developer II | EXERCISEExtra Credit

Do:1 Launch QlikView and save a working copy of the AdditionalExer-

cises.qvw QlikView file in the extra cradit folder for this chapter (C:\QlikViewTraining\DeveloperII\Chapter05\ExtraCredit_Chapter05).

2 To clean up the script a little more, Join the Categories table with the Products table. Make sure not to get any Categories that do not exist in the Products table.

//************** Categories table **************Categories:LEFT JOIN (Products)LOAD CategoryID,

CategoryName,Description AS CategoryDescription,IF(CategoryID = 5 OR CategoryID = 6,

'Footwear','Clothing') AS CategoryType;

SQL SELECT *FROM Categories;

3 Create a pivot table with CategoryType and CategoryName as dimen-sions.

4 Create the following four expressions:

Sales Sum(LineSalesAmount)

COGS Sum (CostOfGoodsSold)

Margin Sum (Margin)

Margin % Sum (Margin)/ Sum (LineSalesAmount)5 Format the table the way you want to.

Figure 1. The resulting pivot table.49

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTS6 ADVANCED CALCULATIONS IN SHEET OBJECTS

6.1 Set AnalysisQlikView has always been good at calculating aggregates for the current selection of data. However, when you wanted to compare results for differ-ent selections in the same chart, you needed to either prepare data in the script or resort to rather complicated expressions with if clauses.

Set analysis changes all that, by making it possible to modify any aggrega-tion function with an arbitrary selection set.

The set may be defined as a bookmark, as an on-the-fly selection in one or more fields, as a function of current selections, the inverse of current selec-tions, previous selections or all data.

The possibilities are endless and yet the syntax is fairly simple and straight-forward.

Indirect Set AnalysisSelections in a field can be stated based on selections in another field, such as

Select all possible values in Customers based on Sales last year.

6.1.1 Overview

Sets can be used in aggregation functions. Aggregation functions normally aggregate over the set of possible records defined by the current selection. But an alternative set of records can be defined by a set expression. Hence, a set is conceptually similar to a selection.

6.1.2 Set Identifiers

There is one constant that can be used to denote a record set; 1. It represents the full set of all the records in the document.

Objectives Introduce advanced calculations in charts and tables, including:

Set Analysis Dollar-Sign Expansion AGGR Function

Complete exercises using examples of each of these functions

Note: A set expression is always enclosed in curly brackets when used, e.g.{BM01}.51

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTSThe $ sign represents the records of the current selection. The set expression {$} is, therefore, the equivalent of not stating a set expression at all. {1-$} is all the more interesting as it defines the inverse of the current selection, that is, everything that the current selection excludes.

Selections from the Back/Forward stack can be used as set identifiers, by use of the dollar symbol: $1 represents the previous selection and is equivalent to pressing the Back button. Similarly, $_1 represents one step forward and is equivalent to pressing the Forward button. Any unsigned integer can be used in the Back and Forward notations. $0 represents the current selection.

Finally, bookmarks can be used as set identifiers. Either the bookmark ID or the bookmark name can be used, BM01 or MyBookmark.

6.1.3 Set Operators

Several operators are used in set expressions. All set operators use sets as operands, as described above, and return a set as result. The operators are as follows:

+ Union. This binary operation returns a set consisting of the records that belong to any of the two set operands.

Exclusion. This binary operation returns a set of the records that belong to the first but not the other of the two set operands. Also, when used as a unary operator, it returns the complement set.

* Intersection. This binary operation returns a set consisting of the records that belong to both of the two set operands.

/ Symmetric difference (XOR). This binary operation returns a set consisting of the records that belong to either, but not both of the two set operands.

The order of precedence is

1 Unary minus (complement)2 Intersection and Symmetric difference3 Union and Exclusion.

Within a group, the expression is evaluated left to right. Alternative orders can be defined by standard brackets, which may be necessary since the set operators do not commute, i.e. A + (B C) is different from (A + B) C which in turn is different from (A C) + B.

Set Operator Examples:

sum( {1-$} Sales ) returns the sales for everything excluded by the current selection.52

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTSsum( {$*BM01} Sales )returns the sales for the intersection between the current selection and bookmark BM01.

sum( {-($+BM01)} Sales )returns the sales excluded by current selection and bookmark BM01.

6.1.4 Set Modifiers

A set can be modified by making an additional or a changed selection.

Such a modification can be written in the set expression.

The modifier consists of one or several field names, each followed by a selec-tion that should be made on the field, all enclosed by < and > as in

Field names and field values can be quoted as usual, e.g. .

There are several ways to define the selection:

A simple case is a selection based on the selected values of another field, e.g. . This modifier will take the selected values from DeliveryDate and apply those as a selection on OrderDate.

The most common case, however, is a selection based on a field value list enclosed in curly brackets, the values separated by commas, e.g. . The curly brackets here define an element set, where the ele-ments can be either field values or searches of field values.

A search is always defined by the use of double quotes, e.g. will select all ingredients including the string garlic.

Note: The use of set operators in combination with basic aggregation expressions involving fields from multiple QlikView tables may cause unpredictable results and should be avoided. E.g. if Quantity and Price are fields from different tables, then the expression sum($*BM01} Quantity * Price) should be avoided.

Note: If there are many distinct values more than a couple of hundred avoid this operation because it is CPU intensive.53

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTSFurther, the selection within a field can be defined using set operators and several element sets, such as with modifier

which will select all years beginning with 20 in addition to 1997, except for 2000.

The above notation defines new selections, disregarding the current selec-tion in the field. However, if you want to base your selection on the current selection in the field and add field values, e.g. you may want a modifier . A short and equivalent way to write this is , i.e. the assignment operator implicitly defines a union.

Also implicit intersections, exclusions and symmetric differences can be defined using *=, = and /=.

Finally, for fields in and-mode, there is also the possibility of forced exclu-sion. If you want to force exclusion of specific field values, you will need to use ~ in front of the field name.

6.2 Dollar-Sign ExpansionDollar-sign expansions are definitions of text replacements used in the script or in expressions. This process is known as expansion - even if the new text is shorter. The replacement is made just before the script statement or the

Note: Searches are case-insensitive and are made over excluded values too.

Tip: Empty element sets, either explicitly e.g. or implicitly e.g. (a search with no hits) mean no product, i.e. it will result in a set of records that are not associated with any product.

Note: A set modifier can be used on a set identifier or on its own. It can-not be used on a set expression. When used on a set identifier, the modi-fier must be written immediately after the set identifier, e.g. {$}. When used on its own, it is interpreted as a modification of the current selection. 54

expression is evaluated. Technically, it is a macro expansion.

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTSA macro expansion always begins with $( and ends with ) and the content between brackets defines how the text replacement will be done. To avoid confusion with script macros we will henceforth refer to macro expansions as dollar-sign expansions.

6.2.1 Dollar-sign Expansion using a variable

When using a variable for text replacement in the script or in an expression, the syntax

$ (variablename)

is used. $(variablename) expands to the value in variablename. If variable-name does not exist the expansion will be the empty string.For numeric variable expansions, the syntax

$ (#variablename)

is used. $(#variablename) always yields a legal decimal-point reflection of the numeric value of variablename, possibly with exponential notation (for very large/small numbers). If variablename does not exist or does not con-tain a numeric value, it will be expanded to 0 instead.

6.2.2 Dollar-Sign Expansion with Parameters

Parameters can be used in variable expansions. The variable must then con-tain formal parameters, such as $1, $2, $3 etc. When expanding the vari-able, the parameters should be stated in a comma separated list.

If the number of formal parameters exceeds the number of actual parame-ters, only the formal parameters corresponding to actual parameters will be expanded. If the number of actual parameters exceeds the number of formal parameters, the superfluous actual parameters will be ignored.

The parameter $0 returns the number of parameters actually passed by a call.

6.2.3 Dollar-Sign Expansion with an Expression

Expressions can be used in dollar-sign expansions. The content between the brackets must then start with an equal sign:

$(=expression)

The expression will be evaluated and the value will be used in the expan-sion.

Example:

Note: Macro expansion is unrelated to script macros (VB or Java script defined in the script module).55

QlikView Developer II | 6 ADVANCED CALCULATIONS IN SHEET OBJECTS$(=Year(Today())) returns the calendar year based on the system date, so if your system date is 28 May 2009, 2009 would be returned

$(=Only(Year)-1)returns the year before the selected one

6.3 AGGR FunctionAGGR is a powerful QlikView function that returns a set of values of expression calculated over dimensions. The result can be compared to the expression column of a local chart, evaluated in the context where the aggr function resides. Each dimension must be a single field and cannot be an expression (e.g. a calculated dimension).

If the expression argument is preceded by the nodistinct qualifier, each com-bination of dimension values may generate more than one return value, depending on underlying data structure. If the expression argument is pre-ceded by the distinct qualifier, or if no qualifier is used at all, each combina-tion of dimension values will generate only one return value.

By default, the aggr function will aggregate over the set of possible records defined by the selection. An alternative set of records can be defined by a set expression.

By using this function in calculated dimensions it is possible to achieve nested chart aggregation in multiple levels.

When aggr is used in chart expressions it is possible to achieve sum-of-rows totals in a pivot table.

Examples:aggr( sum(Sales), Country ) aggr( nodistinct sum(Sales), Country ) aggr( sum(Sales), Country, Region ) count( aggr( sum(Sales), Country )) 56

QlikView Developer II | EXERCISEAdvanced Calculations Exercises

Do: Set Analysis ExerciseCreate a Straight Table chart that displays a comparison of annual sales by CompanyName based on the year selected by the user.

1 Navigate to the c:\QlikViewTraining\DeveloperII\Chapter06 directory and open the Set_Analysis.qvw file.

2 Save a copy of the file to preserve the original in case you want to start again from the beginning later. Do this by using the File | Save As com-mand. There is also a QlikView file ending in _Solution.qvw contain-ing the completed exercise for your reference.

3 Double check to be sure there is a list box on the sheet for the Year. If not, add one.

4 Right-click in a blank area of the sheet and New Sheet Object | Chart from the context menu.

5 Click on the Straight Table icon (the lower right corner of the Chart Types) and type Annual Comparison in the Window Title. Click on the Next button.

6 Add Customer to the Used Dimensions and click Next7 Create the following three Expressions using the Labels provided:

8 Click Finish9 Save your QlikView file and then continue to edit the Annual Compari-

son straight table.10 Set the Sort order to match the depiction, below, remembering that Cus-

tomer should be set to Text

Label Expression1 =Only(Year) Sum({$} LineSalesAmount) 2 =Only(Year)-1 Sum({$} LineSalesAmount)3 =Only(Year) & ' vs '

& (Only(Year)-1) Sum({$} LineSalesAmount) - Sum({$} LineSalesAmount)57

QlikView Developer II | EXERCISE11 On the Visual Cues tab, make the negative values for the year-to-year comparison red and the positive values green.58

QlikView Developer II | EXERCISE12 Return to the General tab and add a Calculation Condition to ensure that the user selects a Year to begin the comparison by entering the fol-lowing into the Calculation Condition box

Count(distinct [Year])=113 Click on the Error Messages button on the General tab and then on Cal-

culation Condition Unfulfilled in the Standard Messages list.14 Type: Select a Year to compare with a previous year in the Custom Mes-

sage box and click OK.

15 Click OK again to close the chart properties dialog.59

QlikView Developer II | EXERCISE16 With 2009 selected in the Year list box you added at the beginning of the Exercise, your straight table should look something like the one below:

17 Save your work.

Do: Advanced Set Analysis1 Continue working in the file you have been using in this exercise chapter

so far. 2 Create a chart that compares sales of products in the category Babywear

with products in the category Childrens Wear over time for the Nice sales office.

3 To do this you will need to create an expression using Set analysis and $ Expansion instead of traditional if() statements.

4 Set the category name to find Babywear and the office to be 4. 5 Pay attention to the and {}.60

QlikView Developer II | EXERCISEDo: AGGR Function Exercise1 Continue working in the file you have been using in this exercise chapter

so far. 2 Create a table that shows if there is any link between the number of

orders placed by customers and the average order value. The table should provide information on how many customers have placed one order, two orders etc, and also the average order value.

3 There are three steps to this process. First, create a calculated dimension for the number of orders (as in how many customers had one order, two orders, three orders, etc.). This requires aggr. First, to count the number of orders, use

count(distinct OrderID)and then to aggregate those against the Customer Dimension use

aggr(......, CustomerID)Putting it together,

aggr(count(Distinct OrderID), CustomerID)will create the necessary order "buckets" based on customer

Solution:

sum({$} LineSalesAmount)

sum({$} LineSalesA-mount)61

QlikView Developer II | EXERCISE4 Next you need a count of customers to populate the # of order buck-ets we created in the first step.

Count(distinct CustomerID)5 And, finally, we need to find the average order amount.

Sales is from the Sales Detail Table and is a line item for every product sold. Thus one order could have several products and thus several lines with Sales data, so we need to aggregate by OrderID to get the total sales amount for an order

aggr(sum(LineSalesAmount),OrderID) gets you that number and adding the avg() gives you the requested aver-age order amount

avg(aggr(sum(LineSalesAmount),OrderID))62

QlikView Developer II | 7 SCRIPTING AND DATA MODELING CHALLENGES7 SCRIPTING AND DATA MODELING CHALLENGES

7.1 Link Tables and Concatenated TablesLink tables combine information that has different levels of detail or fre-quency. Generally, they are best used when the metrics being measured do not have the same base of time, perhaps one is an aggregation over a time period, i.e., by Month. Link tables are also used when data does not meet the business requirements of a calculation. So you can think of them in terms of their name. They are useful to link information that would not oth-erwise be automatically linked with QlikViews associative logic.

In other words, when loading data, certain scenarios can occur whereby you need to store fact type information in different tables, such as transactions separated from budgets. Each type of fact may need to join with multiple common dimensions (such as Time and Product) which would then cause loops in the datastructure. To circumvent such scenarios, we create linking tables that store the individual ways that each fact table joins to the common dimensions.

7.1.1 Concatenation

Concatenation can be used to reduce the number of fact tables in a structure if there is sufficient commonality between the fact tables to validate such a combination. This can make a data model more efficient.

When fact tables contain the same granularity of information and a com-mon key is found between both facts, an outer join operation is preferable to a concatenate operation, as the output records would be linked on a record-by-record basis and hence single record operation will be possible. (e.g. For a given SalesAndBudget fact record, to be able to calculate the dif-ference between the sales amount and the budget amount.)

7.2 Calculating net change within a fieldWhile loading data, you can look at what is being loaded using the peek() function and use it in calculations that express the relationships between records. This can be useful for understanding an account balance, over time, for example.

Objectives Understand the use link tables Calculate net change within a field Use advanced functions for handling time Know why dynamic aggregation and interval matching are important 63

QlikView Developer II | 7 SCRIPTING AND DATA MODELING CHALLENGESWe can also see what has been loaded in the load process with the previ-ous() function. The previous() function can only be used to look one record back in the load process. If you need to look further back than one record in the load process, you have to nest the previous() function as in:

previous(previous(myRecord))This can be problematic and cumbersome when you want to see the total net change or percent change.

Let us assume we need to count the number of units that are at a given site within a week and need to find the percent of utilization for a unit at a given site.

If the units are at multiple sites but return to the same site within the same week, peek() and previous() can work but require complex coding with FOR loops or RESIDENT loads. However, we can get around that by grouping the information from a tracking table and Left Joining the result back into the tracking table.

The steps in the exercise demonstrate the difference.

7.3 Functions InDate, Data & DateIslandsThis section includes concepts used to create documents: InDate and Data/DateIslands.

Data/DateIslands are tables that are not directly linked with other tables (tables that are independent of any explicit links). You can think of this as an array of values that can change with a selection by the user and will force new results to be displayed in the Qlikview Data Model via a non-key field connection.

However, when using this approach you will not change the selection state in the underlying QlikView Data Model. This is usually used for KPIs so that, when moving to a tab with more detail, the user has a full view and selection of data for filtering.

If there is a selection made in the QlikView Data Model outside of a Data/DateIsland, that select will remove data used in the calculation of a KPI and could result in blank or misleading KPI displays.

The InDate() functions are used in conjunction with the Data/DateIsland.

7.4 Aggr() Note: If you have already completed Chpater 6, skip this explanation. Aggr() is an advanced function that allows you to calculate an aggregation of multiple dimensions. As a rule, you can not aggregate an aggregated 64

QlikView Developer II | 7 SCRIPTING AND DATA MODELING CHALLENGESexpression (nested aggregations). The exception to this restriction is to use aggr(); the format is:

aggr([distinct | nodistinct] expression {,dimension})

The best way to think of aggr() is that, by using this function, you can cre-ate a mini QlikView data model on the fly, without needing to edit your data load script or reload from the source data for your QlikView file. It allows for nested aggregation. You can add more dimensions than are represented in charts such as a Pivot Table.

Aggr() is used in the user interface, rather than in the load script. Typically, it is used in a chart object expression, as we saw in the previous chapter.

7.5 Class() Class() is an aggregation function that can be used to create buckets of information similar to what you would create in an Accounts Receivable document showing aged accounts. Another way of thinking of Class() is that you can group your dimension values.

7.6 Dynamic AggregationIn QlikView, you can assign a field value to a variable. Subsequently, you can use the value that the variable holds in an expression. This is similar to a pointer in C-programming or other similar programming languages.

QlikView calculates information on the fly. With this approach and using a table to hold expressions, you can create a reference to the formulas located in the table which, in turn, effectively creates a library of expressions that can then be used to dynamically change your display of information.

The user will be able to cycle through the KPIs by selecting choices in a List Box.

7.7 IntervalMatchThe IntervalMatch script statement facilitates the mapping of dates to peri-ods or records to slowly changing dimensions, useful for creating a fully functioning data model across the appropriate time dimensions in your busi-ness requirements. The implementation of IntervalMatch requires a few additional steps than simply to apply the function if we are to avoid syn-thetic keys being created in our data structure.

IntervalMatch() is a script statement that has the same functionally as BETWEEN used in a SQL statement. There is one downside to IntervalMatch(), it creates a synthetic table ($Syn table). $Syn tables can have a high cost in 65

QlikView Developer II | 7 SCRIPTING AND DATA MODELING CHALLENGESmemory and User Interface performance can suffer. One solution to explore is to LEFT JOIN the IntervalMatch() table into a parent table. Here is a small example of IntervalMatch(). We are maintaining the approach of keeping the examples small so that they can be easily verified.

In this exercise we will use IntervalMatch() to find the hours that an employee is working on-site.66

QlikView Developer II | EXERCISEExercises: Advanced Scripting and Design

Do: Link Tables Exercise1 Launch QlikView and save a working copy of the QlikView file in the

working directory for this chapter and section (C:\QlikViewTrain-ing\DeveloperII\Chapter07\LinkTables\Linktable.qvw).

2 Open the LinkTable.XLS Excel file and examine the Sales as well as the Budget fields. You will see that four customers have a Budget and four customers purchase goods each month but not the identical customers that have a Budget. Now reload your working copy of the LinkTa-ble.qvw file. All the information looks correct. Go to the Table Viewer and look at the structure of the Data Model. A Synthetic Key and a Syn-thetic Table have been created. This can add processing overhead to the document and, in some rare cases, can return unexpected results.

3 Comment the script on the SynKey tab, uncomment the script on the Using a Key tab and reload the document. Note that we are trying elim-inate the Synthetic Key by concatenating the fields that are causing the Synthetic table. Check the result from the QVW and compare them to the Excel sheet. You can see that we are returning incorrect information because of key field mismatches; customers are missing from each table.

4 Comment the script on the Using a Key tab and uncomment the script on the Link Table tab and reload the document.Here we are loading the information from the Sales and Budget table two times. This generates an intermediate table that has common infor-mation from both the tables. The level of detail is maintained in separate tables for both the Sales and Budget. The link table must be loaded Dis-tinct (only unique key combinations) and it will establish correct rela-tionships.

5 Create a Table Box that includes all the columns.67

QlikView Developer II | EXERCISE6 Comment the script from step 3 (on the Link Table tab) and uncomment the script on the Concatenate tab. Reload the document. Note the use of a RecSource field to differentiate concatenated records. Note also the use of explicit references to CONCATENATE and the table into which the concatenation will happen. Finally, observe the LEFT JOIN that adds the Customer field to the table and note that this results in a Data Model with a single table. Suggest an alternative way to accomplish the same result.

7 Look at the result. You will see the same results as Step 1 and Step 3.Notice that there are null values. Now if you sum Sales Amount and Budget Amount, they will balance.

Note: This is similar to a union in SQL. Null values do not have a high cost in QlikView as they normally do in SQL.

Remember: Link tables resolve differences in granularity between fact tables joining to the same dimensions.68

QlikView Developer II | EXERCISEDo: Join Exercise1 Open the LeftJoin.QVW file located in the LeftJoin folder, save a work-

ing copy and run the script on the Previous tab. Create two List Boxes and use the fields Units and Dept.

2 Select the Unit numbered 54543. Create a Table Box using all fields and look at the counts.

3 We still dont have a full count for the units at Dept A2. We would have to reload the table again. But where do we stop?

4 The solution is to aggregate the data using a GROUP BY clause from the Tracking table data and LEFT JOIN the result to the Tracking table.

5 Comment the information on the Previous tab, uncomment the script on the Left Join tab, and reload the document. With this approach you now can find the total number of times a Unit is utilized across all Depts in any Week. This was done with one additional RESIDENT load.

6 Your result should look like this:69

QlikView Developer II | EXERCISENow you can calculate the percent of utilization of a Unit in a Dept for a given Week.

Remember: Group by statements in combination with left join operations allow for aggregation of values living in tables low down in a parent-child hierarchy and appending to the higher tables, such as summing the value of individual order line items into a single value per order.

Do: Date Island Exercise1 Open the file DateIsland.qvw in the Date_Island folder, save a working

copy of the file and add the following code in the Calendar tab below the MasterCalendar: line in the script.

DateIsland:LOAD DISTINCT

monthstart(OrderDate) AS DateIsland,MonthYear AS DateMonthYear

RESIDENT

MasterCalendar;

2 Reload the script.3 Create a List Box and a Pivot Table. In the Pivot Table use Country and

any CategoryType. For the List Box show DateMonthYear. Use two of the three expressions, below (hint: one of them violates a best practice):

sum(LineSalesAmount * InMonth (OrderDate,DateIs-land,0) *-1))sum(if(InMonth (OrderDate,DateIsland,0),LineSalesA-mount,0))sum(LineSalesAmount)

4 Select a single DateMonthYear. The result should look like this:

Note: The ORDER BY clause is required here, but it will only work for a RESIDENT LOAD or in a SQL SELECT statement. ORDER BY will not work with a LOAD FROM statement.70

QlikView Developer II | EXERCISE5 Last, you need to set a property of the List Box to always have only 1 value selected. Note you cannot do this unless you selected a single value in the step above.

6 Now select various DateMonthYear values and you will see Sales change in one column but not the other.

Remember these key points: Filtering can be achieved without affecting the rest of the data structure. This is achieved through the use of conditional statements and present-

ing the selection criteria of such filters in fields that do not link to the data structure.

Inter-record functions can be used to calculate rolling averages and per-forming period comparison in straight and pivot tables.71

QlikView Developer II | EXERCISEDo: Class () Exercise1 Open the file, Class.QVW in the Class folder and save a working copy of

the file. 2 Create variables called vDate (set vDate to a current date and format

vDate as a date, not a numeric) and vWidth (using a pre-pended v or var to denote variables is a best practice) on the Variables tab under the Set-tings | Document Properties menu

3 Create a Calendar Object, click the Variable(s) radio button, and choose vDate as the variable assigned to the Calendar Object.

4 Create an Input Box and choose vWidth as the only Displayed Variable.

5 Next create a Pivot Table with CompanyName as the dimension and sum(LineSalesAmount) as the expression

6 Last, we will create a calculated dimension in the Pivot Table.

Add the class() function using the following: =class(num(OrderDate - $(vDate)),$(vWidth),'Dt')

Figure 1. Variable settings

Figure 2. Calendar and Input Box Objects72

QlikView Developer II | EXERCISEHere we used 2 variables to dynamically change the start period as well as the buckets that the sales or AR information falls in. The screenshots below show two different vWidth values.73

QlikView Developer II | EXERCISEDo: Dynamic Aggregation Exercise1 Open the file located in the DynamicAggregation folder called Dynamic

Aggregation.QVW, save a working copy and reload it. Next create a pivot table chart that looks like

Where the expression is sum(LineSalesAmount)74

QlikView Developer II | EXERCISE2 When you have finished the chart, edit the script creating the following INLINE table and reload. (Note this could also be done by creating and loading an external text or Excel file a best practice).

3 Next create a variable vMyAgg with the value, =Formula as show below. Remember the equal sign:

4 Create a copy of the first Chart and replace the expression sum(Line-SalesAmount) with

if(Desc = 'Count', num($(vMyAgg),'###0'),num($(vMyAgg),'$###.00'))

Note the test to determine how to format the numeric display (integers or currency). You might try including the format string with the expres-sion in the INLINE table as an additional field and then change the expression accordingly.

5 Add two List Boxes with the columns Desc and Formula plus a Text Box with =vMyAgg in the text area.

6 The result is that you can change your KPIs on the fly. Your result should look as follows: 75

QlikView Developer II | EXERCISE7 This technique can be used to change the presentation of entire tables and pages by changing the Units of Measure or the calculations done within the Sheet Object (as above). It can be used to Show/Hide tables and Charts based on selections made in a List Box. You can create an external file that is a library of expressions and then load the file and create variables that contain your expressions.

Do: Interval Match Exercise1 Open the file IntervalMatch.QVW and save a working copy.2 Within your working copy, create a table showing the sum of hours a

person works and a table of the count of hours worked by each Employee (Emp).

Notice that the line before IntervalMatch() is commented.3 Look at the result in the Table Viewer. You will see that this script has

created a $Syn table. As we learned earlier, a $Syn table may have a high cost.76

QlikView Developer II | EXERCISE4 Uncomment the statement before the IntervalMatch() and reload the document.

Result:

The $Syn table has been removed with an INNER JOIN into the table that has the start and stop values (TabB). This can reduce the over-head of a $Syn table. You also have the discrete records for all occur-rences and can use this table to JOIN to other tables.Note that you have increased the size of TabB. The rows in TabB will increase to the number of rows that would be in the $Syn table.77

QlikView Developer II | EXERCISEExtra Credit

Do: Additional AGGR Exercise1 Open the file, Aggr.QVW in the Aggr folder and save a working copy of

the file.2 In this example we are going to find the average count of Orders by

Country, Company, Year and Month using the aggr() function.

avg(aggr(count(Distinct OrderID),Country,CompanyName, Year, Month))

3 The average LineSalesAmount by Country, Company, Year and Quarter with the expression below:

avg(aggr(sum(LineSalesAmount),Country,CompanyName, Year, Quarter))

4 And, finally, the Total Sales

sum(LineSalesAmount)5 Create the chart shown below using the expressions above.

6 Once you have created the chart, investigate further by changing the time dimension in the dimensional portion of the aggr() function. Determine how many Orders were placed in each Month and how many Months had Orders in them. Try to explain some of the resultant aver-ages in each column.79

QlikView Developer II | 8 DATA MODEL OPTIMIZATION8 DATA MODEL OPTIMIZATION

8.1 Performance Tuning

8.1.1 What makes up the size of a QlikView document? Number of rows Number of columns Number of distinct values / column Data structure Number of sheets / sheet objects

Number of rows Question: Is it really necessary to have 10 years of data in the same

QlikView document? Question: If there are too many detail rows, is it feasible to create a roll-

up version at a higher aggregate level?

Number of Columns

Be selective about what fields are included in the QlikView document Focus on the larger fact tables first But dont ignore the dimension tables

Eliminate fields that are not currently being used All fields might be used in the future but are not used now The goal is to have a Lean QlikView document

Number of Distinct Values / Column

Fields with highly distinct values use the most space

Many of them can be eliminated / truncated altogether System keys Timestamps with minutes / seconds

Others can be transformed by breaking them up into separate components Phone Numbers Timestamps

Objectives Discuss performance tuning Explore the impacts of appropriate document design Introduce security concepts Review best practices81

QlikView Developer II | 8 DATA MODEL OPTIMIZATIONData Structure

Key Fields

Denormalization vs. Normalization

Normalization Definition (Wikipedia):Normalization splits up data to avoid redundancy (duplication) by mov-ing commonly repeating groups of data into a new table. Normalization therefore tends to increase the number of tables that need to be joined in order to perform a given query, but reduces the space required to hold the data and the number of places where it needs to be updated if the data changes.

Number of sheets / sheet objects

Remove hidden sheets and sheet objects that are not being used

Remove unused variables

Remember to keep your QlikView document as trim as possible

8.1.2 Document Design

Avoid Show Frequency

Forcing Proper Object Display Calculation Conditions Show Conditional

Avoid too many active objects on the same sheet

Minimized objects consume no resources, use them whenever appropriate

8.1.3 Security Impacts - Making the Right Choices

QlikView Publisher Security

Section Access

Publisher Security

Breaking up one large QlikView file into multiple smaller files based on row level security

Effective for memory management if security profiles do NOT contain much overlapping data

Section Access

Dynamically reduces the QlikView file at logon based on user authorization82

QlikView Developer II | 8 DATA MODEL OPTIMIZATION8.2 Best Practices for QlikView File Optimization

8.2.1 Complex Dimensions and Expressions Many dimensions and expression that are to be placed in charts or tables require some degree of complex scripting such as IF THEN ELSE statements or WHERE [FIELD1] IS NULL.

Many developers build their initial data model using the processes above but then stop amending the script and work solely within the dashboard/GUI. When designing/writing your script you should already be aware of some of the measures that you are looking to create in the end.

Where ever possible you should look to place all complex formulas and statements within the script of the document and not in the actual dash-board/document objects.

If you are to use complex expressions within a dimension or calculation then you should move as much of this as possible into the scripting of the QVW.

8.2.2 Resource Intensive Expressions

Count ( Distinct 'FieldName')

Replace the count() with sum() and the distinct qualifier by assigning the value '1' to each distinct occurrence as it is read in the script

If ( Condition(Text),..)

Map Text to numeric e.g. by using autonumber and/or do the test in the script

Sum ( If (Condition, 'FieldName'))

Here the aggregation is independent of the table dimensions and the result is distributed over the dimensions of the table. The problem can be treated either by doing the test in the script and aggregating in the table or by doing the whole operation in the script. There are numerous techniques for this e.g. interval match, group by, peek, if.thenelse.

If ( Condition, Sum('FieldName')..)

Included here to emphasize the difference to above. This aggregation is com-pletely contextual

If ( Condition1, Sum('FieldName'), If (Condition2, Sum('FieldName')

The logic of nested If..then else.. is conceptually easy but can often become troublesome to administer. We have seen cases with hundreds of nesting lev-83

els. This will be memory and CPU intensive. Often the "Conditions" can be

QlikView Developer II | 8 DATA MODEL OPTIMIZATIONreplaced by transforming them. A typical example is aggregating quan-tity*price where price is variable. This can be handled by "extended interval match". If two conditions, e.g. " A AND B " are to be satisfied the test might be replaced by a condition "C".

Sort text

QlikView automatically evaluates if a Field is to be treated as numeric, text or general. Fields evaluated as text will be sorted as text which is the slowest sort operation. This can be replaced manually to sort by load order.

Dynamic captions and text objects Expressions can be entered almost anywhere that you can enter text. The evaluation of an expression is however dependent on its environment. Expressions in charts and straight-and pivot-tables that are defined in the expressions dialog are embedded and only calculated when the object is active. For instance they are not calculated when the object is minimized.

On the other hand if the object title is calculated this calculation is per-formed every time any change occurs. We also have numerous ways of defining show conditions, calculation conditions etc. These tests will also be performed at all times. Some expressions are more expensive than others and of course become more expensive the more frequently they have to be evaluated. The introduction of asynchronous calculation has shifted the behavior and these effects may have become more noticeable in your docu-ments. The time functions e.g. Now(), Today() will be evaluated whenever a recalculation has to be done. Especially the Now() function can become quite costly since it causes a recalculation of the document every second. For example

If ( ReloadTime()+3>Now(), 'Old Data', 'New Data')

Here one might consider

If ( ReloadTime()+3>Today(), 'Old Data', 'New Data')

As a simple test, put the expressions into textboxes. Then try sizing the text-box with Now() in it.

8.2.3 Adding Aggregatable Columns in to Your Script

It is sometimes good to add manual columns in to your script to give your document the ability to sum up over these values. This will mean you can then place the complex statement (If) in to your script and have a simple sum in your dashboard object.

To do this you should simply place an IF THEN ELSE statement in to your script that substitutes a database column or a 1 or 0 as a value to enable a 84

summing/count to take place.

QlikView Developer II | 8 DATA MODEL OPTIMIZATIONExamples:

IF(ACTIVE='Y',1,0) IF(ACTIVE='Y',sales_amount,0)

Forcing through a 0 then takes out all of the unnecessary / unwanted values when you apply your SUM.

8.3 Creating Incremental LoadsAn incremental load is a common database task. Incremental loads are used when loading only the new or changed records from a database. QVD files are used to do incremental loads in QlikView. Properly engineered incre-mental loads can significantly improve performance.85

QlikView Developer II | EXERCISEExercises

Do:In this exercise, we will tie together several Qlikview concepts, including:

QVD incremental loading Binary Loading

Your production system is overloaded when QlikView pulls all history. The CIO comes to you and asks that this stop now. How do you fix it?

1 Navigate to the course material folder for this chapter, C:\QlikViewTraining\DeveloperII\Chapter08

2 Open the file in the QVDs folder called BaseLineDataExtraction.QVW reload the file and then choose File | Save As and save the file as Base-LineDataModel.QVW in the QVDs folder.

3 Note that you now have a QVW that contains your Data Model but only has data through 2006 for Orders, OrderDetails, and Shipments. Also note that no Master Calendar has been built yet.

Binary Load

1 Close any open QVWs. Create a new QVW and save it as CreateQVD-fromBinary.QVW in the QVDs folder. Edit the script or click the Qlikview File button in the Script Editor to add the BINARY load of the BaseLineDataModel.QVW file you just saved in