Skip to main content
Announcements
Qlik Introduces a New Era of Visualization! READ ALL ABOUT IT
cancel
Showing results for 
Search instead for 
Did you mean: 
MEJ
Contributor III
Contributor III

Is it possible to load QVD-files using wildcard search?

I wonder if it is possible to have a loadscript where I know the start of the filename, but not the ending.

For example:

Backlog_2019-05-20.qvd

Backlog_2019-05-21.qvd

Backlog_2019-05-22.qvd

Is it possible to say in the FROM statement Backlog_*.QVD or something like that? 

If I am reading from automatically created QVD files with "today's date", I don't know what dates they were stored beforehand. 

Hope the question is understandable. 

2 Solutions

Accepted Solutions
m_woolf
Master II
Master II

Yes. Just like your example.

View solution in original post

Brett_Bleess
Former Employee
Former Employee

Yeah, that is a good point, and about the only protection with QVD files is to be sure you have known good backups you can fall back to if something does go wrong, so I would say your approach is the right way in this case given the times you are seeing etc. and your use case!  Best of luck with things.

Regards,
Brett

To help users find verified answers, please do not forget to use the "Accept as Solution" button on any post(s) that helped you resolve your problem or question.
I now work a compressed schedule, Tuesday, Wednesday and Thursday, so those will be the days I will reply to any follow-up posts.

View solution in original post

5 Replies
m_woolf
Master II
Master II

Yes. Just like your example.

MEJ
Contributor III
Contributor III
Author

Great! Now it worked. I had some other errors in the load script that is now fixed. At least I learned a new function.

However, I would still like to have everything stored into one file, not as 200 separate QVD files after a one year. Not sure what the load time difference is, loading from 1 large file containing 1 million records vs 200 QVD files containing 5000 records in each file. Probably it is a lot quicker to load from 1 file, as it otherwise needs to access 200 files, which probably takes longer.

Brett_Bleess
Former Employee
Former Employee

I would concur with your thoughts regarding single versus many, as with the many in this case the file operations alone will likely add a decent number of seconds to each round, whereas if you load optimized from the single file, I would expect things to be quite quick, just wanted to confirm you were thinking along the correct lines here.

Cheers,
Brett

To help users find verified answers, please do not forget to use the "Accept as Solution" button on any post(s) that helped you resolve your problem or question.
I now work a compressed schedule, Tuesday, Wednesday and Thursday, so those will be the days I will reply to any follow-up posts.
MEJ
Contributor III
Contributor III
Author

Hi Brett,

That was my initial worry, that it would take a long time to load from multiple files. So far it has generated 20 files or so (~400 kB per file), and it takes less than a second to load all. No problem whatsoever. The advantage of having multiple files, is that I can delete one of the files in case it contains invalid data, which happens once in a while. I would prefer to have a gap in the series, rather than having potential spikes in revenue/margins/quantity that would screw up the graphs. As I understand, is that it is rather difficult to correct incorrect data in a QVD file. In case the data load would take an excess amount of time due to too many files, I would load the multiple QVD files into one QVD file at some point.

Thanks

Brett_Bleess
Former Employee
Former Employee

Yeah, that is a good point, and about the only protection with QVD files is to be sure you have known good backups you can fall back to if something does go wrong, so I would say your approach is the right way in this case given the times you are seeing etc. and your use case!  Best of luck with things.

Regards,
Brett

To help users find verified answers, please do not forget to use the "Accept as Solution" button on any post(s) that helped you resolve your problem or question.
I now work a compressed schedule, Tuesday, Wednesday and Thursday, so those will be the days I will reply to any follow-up posts.