We have a HTML with multiple filters. We are populating the ...

[VisuaLizeFOCUS .]

We have a HTML with multiple filters.

We are populating the filters with FOCUS files and chaining is there.

Problem is the performance when we select one value in a dropdown the browser shows a wait signal and takes lot of time to filter other filters.

How can I improve this performance any new ideas

[Martin Yergeau]

How big are the files (nb rows)

How many combination exist (nb chaining)

How complex are they

Are they indexed on chained fields

[robert fuschetto]

I have this issue when pointing to tables with millions of detail rows. I have, in some cases made a new sql table containing an aggregate of the fields we filter on from these table/s and redirect the parameters to them on the HTML PAGE.

[Martin Yergeau]

Chaining on a table with millions of rows will create lag for sure.

Try creating another table with only distinct possible values on which you want to create the filters. For sure at end it will not be millions of rows.

The problem that you are now facing is that each time you are querying the table to display the value on the second, third, filter, the whole table (millions rows) are read.

[Warren Hinchliffe]

With this issue, always check the performance of the queries.

[VisuaLizeFOCUS .]

The thing is guys we have a RC job that create a Permanent Hold file with format to be Focus. So there is no Database HIT. Ill debug more and will post my analysis.

Thanks for the support guys.

[Martin Yergeau]

There you go !

If its a RC job that creates the hold file you have the possibility to add another file that will only contain the distinct values that will be used in the filters boxes.

Once the permanent hold file is done add a step to read that file and create another one with the key fields needed for the filters that needs to be chained.

You will gain a lot of performance having a distinct table (file) with only values to filtre from.

As per example this is not a large file but it will show you what I mean.

If I extract all record from GGSALES sample table, I will end up with 4317 records.

Then creating filters against that table will always querry those 4317 records.

But if I create an aside table with only the key fields (let say I care about category & product), I end up with only 10 records

So, filters will query against 10 rows.

-* To create a permanent indexed Focus file for filters

TABLE FILE GGSALES

BY CATEGORY

BY PRODUCT

ON TABLE HOLD AS MyFolder/MyFiltersValues FORMAT XFOCUS INDEX CATEGORY PRODUCT

END

[VisuaLizeFOCUS .]

This issue is now resolved guys.

Solution:

 

We created a subset of that FACT Permanent Hold file having 2.6M records.

The Subset contains combination of columns used in the filter for cascading.

Since the HTML was built on WebFOCUS 7 version we created a new one in 8202m version.

Performance of HTML page and filters is now good.

 

Thanks for your help Guys

[robert fuschetto]

we do this often. On the HTML page I point parameters to the keyed file (an aggregate of required parameters that is far smaller) to collect selection that gets fed to the conditions against the main file.

Alas, the move to Designer doies not allow such behviour. The filters, so far, only seem to be able to come from the main file being accessed. So I am back to slow chaining for some stuff.

[Martin Yergeau]

I have not used DS yet, but I think that you should open a case and ask the question regarding filter against another file than the main one