Document Analyzer V2 -- a major update

[rwunderlich]

I've posted a major update "V2" to the Document Analyzer tool. It is available for download in the Tools section of QlikviewCookbook.com here Document Analyzer.

The primary function of Document Analyzer remains to identify unused fields that can be removed from your model. It has been expanded to assist in identifying slow charts.

There are a number of improvements over the 1.15 version, including bug fixes, a new UI, object calc times and improved analysis of variable and expression usage.

An important change is that only qvws that have been saved in QV10SR2 or later may be analyzed with this version. If you need to analyze an older document, save it first in QV10.

if a prj directory exists, it will be leveraged for the analysis. This can speed up the analysis run for documents with a lot of objects.

Use this thread if you have comments or questions. Enjoy.

-Rob

[rwunderlich]

Can you contact me via the form at Contact | Qlikview Cookbook? I'd like to get a look at the document analyzer run and try to resolve this.

Thanks,

Rob

[Anonymous]

Yes thank I will first thing tomorrow ... I'm on New Zealand time

[Anonymous]

Rob,

Great tool!   I have run into a small glitch.  There are fields in my drop field statement list that we use for several purposes.

1)  I use the RecNo field so that each record is unique.  Often we have transactions that have been voided and re-entered.  QV does not aggr records that are not unique.  So, If I have a sales record, a reversal and then the recreation of the sales record, I only get 2 records showing up in my data objects.  With the RecNo field, I get all 3.  If I don't use RecNo, the two unique records ( original and voided ) cancel each other.  QV ignores the new transaction because all of data in the new record is identical to the first.  I would hate for people to just globally go and drop all of these fields without considering the use for each.

2)     We have fields used in our data ( "IsEligible" for example ).  This field is used when we reduce data for publishing.  It is not used in any objects, but certainly would be painful if removed.

3)     We have fields that are used in Multi-Boxes and List Boxes that the tool is suggesting we drop.

One of the previous comments suggested just copying the drop field list and adding it to a new tab at the end of your loader.  I would caution against such an approach and would recommend an individual assessment of each field prior to dropping.

If you have any ideas why the fields in the list and multi boxes are slipping through the analysis on the tool.  I would sure appreciate your feedback.  I am using QV11SR9.

Mike Czerwonky

[#5]

I have downloaded the doc analyzer to a server environment.  When I try to open it I get a message from QlikView (ver9) that it is unable to load it.  What can I do to correct this?

Alex

[rwunderlich]

Alexander,

Document Analyzer is only supported on QV versions 10 & 11. You can analyze documents that were saved in earlier versions, but Document Analyzer itself must run in QV10+ (V11 recommended).

-Rob

[rwunderlich]

Hi Mike,

Thanks for the useful comments. As noted in the (sparse) notes in DA, the field usage detection is imperfect and has some limitations. You must filter the recommendations with common sense. I even resisted including the 'Drop Fields" listbox because I didn't want people to take it as gospel.

I'm aware that I don't detect field usage in a few situations like:

- Indirect references

- Publisher loop and reduce.

- Server objects (fix is in the pipeline)

I would echo your advice to not just blindly copy the drop field list into your script.

I would like to look at the missing references in the multi boxes and listboxes you mentioned. If you're willing to work with me on that, please contact me at Contact | Qlikview Cookbook

-Rob

[#8]

Hi Rob, I have a application that hangs and can't figure out why.    I am using QV11 SR7 and running DocumentAnalyzer_V2.3.qvw    

All of my other files are working, except this one.

Do you have any suggestions on how to debug this?

[rwunderlich]

Can you describe the "hang" condition a bit more? Is this during the Extract process or the Reload process?

If during the Extract, please post the last 20 lines or so of the DocumentAnalyzer.log. It's in your temp folder, see a few entries back in this thread for location.

If hanging in the Reload, can you paste the end of the Document log?

-Rob

[#10]

It is hanging during the Extract Process.  Here are the DocumentAnalyzer log details.  

writing ObjectInfo for Document\CH336

Writing xml file C:\Users\jcc\AppData\Local\Temp\qvwork\Sales 360 Newest v3\Document\CH336.xml

Extracting Expressions

Extracting Dimensions

Extracting Fonts

Extracting Actions

writing ObjectInfo for Document\TX2297

Writing xml file C:\Users\jcc\AppData\Local\Temp\qvwork\Sales 360 Newest v3\Document\TX2297.xml

Extracting Expressions

Extracting Dimensions

Extracting Fonts

Extracting Actions

writing ObjectInfo for Document\TX2298

Writing xml file C:\Users\jcc\AppData\Local\Temp\qvwork\Sales 360 Newest v3\Document\TX2298.xml

Extracting Expressions

Extracting Dimensions

Extracting Fonts

Extracting Actions

writing ObjectInfo for Document\TX2575

writing ObjectInfo for Document\TX2574

writing ObjectInfo for Document\TX2577

writing ObjectInfo for Document\TX2572

writing ObjectInfo for Document\TX2573

writing ObjectInfo for Document\TX2570

writing ObjectInfo for Document\TX2569

writing ObjectInfo for Document\TX2571

writing ObjectInfo for Document\TX2555

writing ObjectInfo for Document\CH446

Thanks for looking at this.

[rwunderlich]

It looks like it's having a problem with CH446. Can you remove that chart from a copy of the doc and see if the extract runs without it? If so, then we can focus on that chart.

-Rob