How to search a field in qvd file

[mfcelleri]

Hi guys.

I have several qvd files, which differ in the number and name of fields.

But I need to upload all the files with all their fields.

Is there a way to know if a file has a field or not, so that during the loading process, it does not generate an error when reading a file that does not have those certain fields?

Thanks and regards.

[sbaro_bd]

Hi @mfcelleri ,

If you want to load all fields from a QVD, use * in your LOAD statement like :  LOAD *  FROM [lib://path/qvd_name.qvd](qvd).

The star symbol will take all fields avaible in the qvd.

Regards.

[mfcelleri]

Hi @sbaro_bd ; thanks for your reply.

That it's not what I want.

A short example: Table1 has field1, field2, field3; Table2 has field1, field2, field3, field4; both tables has the same info; but Table2 file's structure change since x's date.

Each qvd file is a cut of information on a specific day; so I have to process each file.

How to validate that in Table1 file (the example) it does not have the field field4; I wan't to prevent give me an error when processing the file.

Hope I explain much better.

[pravinboniface]

I'm not very clear with your requirements, but it sounds like you can come up with some logic using the functions QvdFieldName() and QVDNoOfFields().

Also, you can read 1 row from the QVD into a resident table and then use the FieldName() and NoOfFields() to achieve the same thing.

[sbaro_bd]

Hi @mfcelleri ,

Can you tell us what kind of error do you want avoid exactly when you load your QVD ? Please, provide an example with screenshot if possible (synthetic keys? field not found? wrong concatenation?)

Regards.

[marcus_sommer]

Like the others have already hinted you may query the meta-data of the qvd or doing a pre-load of one record. And the field-information could be extracted + checked + prepared (usually assigning the essential information to variables) and appropriate branched.

It's no big magic but it will need some extra steps. If I understood your description right it are always the same data - just a few fields are coming/going over the time respectively depending on different sources/channels and similar stuff. In this case you may shortcut the approach by loading the data with a field-wildcard and enforcing a concatenation of the loads, for example with:

t: load 'dummy' as Dummy autogenerate 1;

for each file in filelist('path/file*.qvd')
   concatenate(t) load * from [file] (qvd);
next

drop fields Dummy from t;