topic Re: Re: loosely coupled tables in QlikView

loosely coupled tables

arulsettu — Sat, 18 Jan 2014 05:44:20 GMT

Hi i am very new to qlikview can any one explain what is the use of loosely coupled tables with example

thanks in advance...

Re: loosely coupled tables

Anonymous — Sat, 18 Jan 2014 05:50:30 GMT

What is loosen table and link table?

Re: loosely coupled tables

Sat, 18 Jan 2014 08:50:58 GMT

hi, check this out!!!

How to "Loosely Couple" a Table?

Re: loosely coupled tables

arulsettu — Mon, 20 Jan 2014 05:55:06 GMT

can any one explain this with simple scenario...

Re: Re: loosely coupled tables

Siva_Sankar — Mon, 20 Jan 2014 06:28:17 GMT

read the "Loosely Coupled Tables" chapter(page 118). attached is the reference manual.

Regards.

Siva

Re: loosely coupled tables

Anonymous — Mon, 20 Jan 2014 15:19:26 GMT

The best way to deal with the loosely coupled tables is to avoid using them.

Re: loosely coupled tables

marcus_malinow — Mon, 20 Jan 2014 16:06:14 GMT

Agreed in most circumstances, however I saw an interesting article the other day about using loosely coupled tables in a security model. Wish I could find it now though...

Re: loosely coupled tables

Anonymous — Mon, 20 Jan 2014 16:18:49 GMT

Marcus Malinow,

I built a quite complex application using this loosely coupled tables many years ago. As a result, client can't use this application in the newer QV versions because the behavior of the loosely coupled tables has changed , and more than once , and in the different ways . So, I'm not touching this feature anymore. Especially now, when QV has set analysis so we can ignore selections in a more reliable way.

Regards,

Michael

Re: Re: loosely coupled tables

Fri, 24 Jan 2014 08:45:14 GMT

arul: do you know what Circular references are? Perhaps you should start there.

Re: Re: loosely coupled tables

Anonymous — Sun, 26 Jan 2014 19:09:43 GMT

Loosely coupled tables are created automatically when the data model(3 or More Tables) that includes circular references is loaded into QlikView,to avoid them the circular references create a loop in the QlikView internal logic. These loosely coupled tables need to be handled in order to visualize data in a way that is expected and understandable.

See Article Circular References

Any table can also be made loosely coupled interactively from this dialog or via macros. Additionally, it is possible to declare loosely coupled tables explicitly in the script via the Loosen Table statement.

The normal QlikView associative logic is disconnected internally for loosely coupled tables. This means that selections in one field do not Associate through to the other fields in the table. its very useful to avoid circular references in the data structure in various scenarios. For more examples of how this feature can be used, please refer to the QlikView Reference Manual - Intentionally Creating Loosely Coupled Tables

One or more QlikView internal data tables can be explicitly declared loosely coupled during script execution by using a Loosen Table statement.

The use of one or more Loosen Table statements in the script will make QlikView disregard any setting of tables as loosely coupled made before the script execution.

The syntax is:

Loosen Table ~~tablename [ , tablename2 ...]~~

Either syntax: Loosen Table and Loosen Tables can be used.

Example:

Table1:

Select * from Trans;

Loosen table Table1;

Note!
Should QlikView find circular references in the data structure which cannot be broken by tables declared loosely coupled interactively or explicitly in the script, one or more additional tables will be forced loosely coupled until no circular references remain. When this happens, the Loop Warning dialog, gives a warning.

Re: loosely coupled tables

hic — Mon, 27 Jan 2014 03:53:18 GMT

I have never seen a use case where the best solution is to use a loosely coupled table. So - don't use them.

QlikView has loosely coupled tables to be be able to cope with the situation of a script that creates a non-logical data model.

HIC

Re: loosely coupled tables

arulsettu — Mon, 27 Jan 2014 07:21:57 GMT

ok....thanks

Re: loosely coupled tables

subhashgherde — Fri, 08 Sep 2017 19:43:59 GMT

When data that includes circular references is loaded into QlikView, loosely coupled tables are created automatically. This prevents that the circular references create a loop in the internal logic. The loosely coupled tables need to be handled in order to visualize data in a way that is expected and understandable.

Re: loosely coupled tables

subhashgherde — Fri, 08 Sep 2017 19:46:41 GMT

Loosely coupled tables

Intentionally Creating Loosely Coupled Tables

In some special situations, you might want to disconnect some data from the normal QlikView logic. You can use loosely coupled tables to restrict selections in one field to propagate through to the other fields in the table.

To set a table to be loosely coupled, open the Document Properties dialog and select the Tables tab.

This chapter includes some examples of how loosely coupled tables alter the QlikView logic.

Each of the following three table boxes represents a table read into QlikView:

If the value 2 is selected in field B the following will happen:

The selection ripples through all tables, and excludes values that are not associated.

Now, let's keep this selection but make Table2 loosely coupled. This means that the logic will be cut between the fields A and C in Table2. The result will look this:

Note that Table2 shown here is a table box and not the table itself. The table box will show all possible combinations between the fields of its columns. Since there is no logic between fields A and C all combinations of their respective possible values are shown.

Example:

The three tables below are found in a rather typical structure: one transaction table and two dimension tables associating to it via one field each.

You can visualize sales per year and product group using a pivot table. Next to the pivot table, two list boxes show the dimension fields:

Even though this is a correct pivot table, the effects of QlikView's logic could now potentially lead to undesired results. If the year 2012 is selected, you get the following result:

Product group Z is no longer visible in pivot table. This is natural, since the value Z in the field ProdGrp has been excluded by the selection of the value 2012 in the field Year. However, when you analyze sales of year 2012, you probably want to see Z in the chart with a 0 in the sum(Amount) column, so that it is clear to everyone that product group Z exists and that nothing has been sold in 2012.

In some sense, the two fields Year and ProdGrp have nothing to do with each other, and therefore should not interact just because they happen to be associated via the Trans table. This can be dealt with by declaring the Trans table as loosely coupled, which changes the layout of the pivot table:

Note that the selection in the Year list box does not render any value in the ProdGrp list box anymore.