Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
This release as serious mafunction regarding usage of "filters" reading text or excel files. If you use transformations, such as conditional suppression of rows, you will retrieve only first row values.
This is the (actual) answer of QT support : go back to 8.50.6261 (SR4) or up to 9.
We've just passed one day installing a new customer site with QVS + QVP (with qve ocx 6299), and the whole data model is a mess...
Meanwhile finding a solution, I advice not to use SR5 in production, cause it may raise problems in well functionning existing applications !
Actual QT position : Registered as a bug (ID 22138), will be fixed in next release scheduled... end of October.
no comment.
Bonjour Michael,
If we only use a filter clause to suppress rows, a work-around is to rewrite script using where clause. We used this workaround for one of our project. But for live application it is different. First, customer has first to notice that his data model changed, and diagnostic that the filter is no more working. Second he has to rewrite part of the script, and run into comparison tests before putting it in production.
But if you use other filters, like duplicate missing column data in example below, this work-around doesn't work, and the only solution is then to downgrade to SR4.... This is precisely the case for our recently installed customer.
LOAD ... FROM
[catalogue.xls]
(
biff,
header is 2 lines,
explicit labels,
table is [Sheet1$],
filters
(
Replace(3, top, StrCnd(null)),
Replace(13, top, StrCnd(null)),
Remove(Row, RowCnd(CellValue, 16, StrCnd(null)))
)
)
;
If we replace the third filter by " WHERE (len(text(@18)) > 0) ", we still need both "replace" filters (we can't change source data), and the bug still occurs.
We have a customer patch available for you regarding bug 22138
Please contact support@qliktech.com for the patch
We are ofcourse very sorry for the problem caused you
Kind regards,
Qliktech Problem Management
This issue has been solved in patch 6303.
However, the official release is still SR5 (6299), which may still cause data corruption into well functionning data models.
This is the reason why this alert should remain active.