Created attachment 152626 [details] 7Z COMPRESSED ANONYMIZED XML SUMMARY UNDER REPORTS/NET WORTH - A REPORT WITH SAVINGS ACCOUNTS AND INVESTMENT ACCOUNT BALANCES BY YEAR, IF DUPLICATED/COPIED, THE NEW VERSION WILL LOSE ALL INVESTMENTS IN ADDITION, NEW NETWORTH REPORTS CREATED TO SHOW AN ANNUAL BALANCE OF A MIX OF SAVINGS AND INVESTMENT ACCOUNTS WILL NO LONGER SHOW INDIVIDUAL INVESTMENTS. MY REPORTS WERE SET-UP PRIOR TO VERSION 5 AND ONLY NOTICED THIS BUG RECENTLY AS I NEEDED TO MODIFY THE REPORTS TO ADD SOME NEW INVESTMENTS, SO UNABLE TO DETERMINE WHEN THIS BUG FIRST MANIFESTED AS I HAVE BEEN UPGRADING WITH NO SIGNIFICANT ISSUES SINCE VERSION 4.6 OR SO. STEPS TO REPRODUCE 1. OPEN ATTACHED FILE 2. UNDER REPORTS/NETWORTH SELECT REPORT R000084 OR R000133 THEN THE NEW BUTTON TO CREATE A COPY 3. VIEW THE NEW COPY OBSERVED RESULT NEW REPORT IS MISSING INVESTMENTS EXPECTED RESULT REPORT SHOULD BE A DUPLICATE WITH THE SAME MIX OF ACCOUNTS AND INVESTMENTS SOFTWARE/OS VERSIONS Windows: 10 ADDITIONAL INFORMATION
(please don't use all caps. it looks like you are shouting. thanks.) Using the latest Appimage Version 5.1.3-7599ae179 on Linux (I don't have a Windows instance handy) the copies of those two reports seem to be identical to the original reports. I don't recall any recent changes to reports, so I don't know if it might be platform dependent, or possibly fixed by some unrelated change. Can you try the latest version from https://binary-factory.kde.org/job/KMyMoney_Release_win64/lastSuccessfulBuild/artifact/ to see if the problem is still present. Also, please report the exact version as shown in Help/About KMyMoney.
It could well be, that the change to fix #416902 causes your problem. It was introduced around 5.1.0.
Git commit a14607ce6fe056a9321595a631dc1eebbf0f5c4a by Thomas Baumgart. Committed on 08/10/2022 at 16:26. Pushed by tbaumgart into branch 'master'. Allow access to investment sub-accounts in all pivot based reports M +1 -1 kmymoney/plugins/views/reports/kreportconfigurationfilterdlg.cpp https://invent.kde.org/office/kmymoney/commit/a14607ce6fe056a9321595a631dc1eebbf0f5c4a
(In reply to Jack from comment #1) > (please don't use all caps. it looks like you are shouting. thanks.) > Using the latest Appimage Version 5.1.3-7599ae179 on Linux (I don't have a > Windows instance handy) the copies of those two reports seem to be identical > to the original reports. I don't recall any recent changes to reports, so > I don't know if it might be platform dependent, or possibly fixed by some > unrelated change. Can you try the latest version from > https://binary-factory.kde.org/job/KMyMoney_Release_win64/ > lastSuccessfulBuild/artifact/ to see if the problem is still present. Also, > please report the exact version as shown in Help/About KMyMoney. Hello, my apologies about the caps, I certainly was not shouting, just a long tiring day and didn't realize the caps was on. I can duplicate the bug on the following versions under windows: 5.1.1-dcc7ca980 - my daily driver for the past ~9 months 5.1.3.-5b-03a2216 - when I wanted to see if it was a bug to the above version I just tried the latest version 5.1.3-8910ed3bb and still see the same behavior with missing investments. Hope this helps.
(In reply to Thomas Baumgart from comment #3) > Git commit a14607ce6fe056a9321595a631dc1eebbf0f5c4a by Thomas Baumgart. > Committed on 08/10/2022 at 16:26. > Pushed by tbaumgart into branch 'master'. > > Allow access to investment sub-accounts in all pivot based reports > > M +1 -1 > kmymoney/plugins/views/reports/kreportconfigurationfilterdlg.cpp > > https://invent.kde.org/office/kmymoney/commit/ > a14607ce6fe056a9321595a631dc1eebbf0f5c4a From the previous comment @2022-10-08 16:24:52 UTC, should I be testing on a version prior to 5.1.0? If so not sure where to download. Otherwise, from this comment, I gather an the next build should have a fix that can be tested? Thank you
Thomas' commit was to master branch, which will not be in a release for a while. However, it looks like the commit can be backported/cherry-picked to the 5.1 branch, in which case it would be in the 5.1.4 release, not yet scheduled, or in a release appimage (https://binary-factory.kde.org/view/AppImage/job/KMyMoney_Release_appimage-centos7/) following the commit. Using a copy of your data file, you could test with an appimage from master branch (https://binary-factory.kde.org/view/AppImage/job/KMyMoney_Nightly_appimage-centos7/), but that is not suggested for production use. Testing with a version prior to 5.1.0 might confirm that the change Thomas referred to is indeed the problem, but it's probably better to wait and test with some version which has the fix.
(In reply to Jack from comment #6) > Thomas' commit was to master branch, which will not be in a release for a > while. However, it looks like the commit can be backported/cherry-picked to > the 5.1 branch, in which case it would be in the 5.1.4 release, not yet > scheduled, or in a release appimage > (https://binary-factory.kde.org/view/AppImage/job/KMyMoney_Release_appimage- > centos7/) following the commit. > Using a copy of your data file, you could test with an appimage from master > branch > (https://binary-factory.kde.org/view/AppImage/job/KMyMoney_Nightly_appimage- > centos7/), but that is not suggested for production use. > Testing with a version prior to 5.1.0 might confirm that the change Thomas > referred to is indeed the problem, but it's probably better to wait and test > with some version which has the fix. Thank you. I'll see if I can boot into a Linux to check it out. I'll keep an eye here to test on Windows as well when ready.
Sorry, I forgot you were on Windows. You can get a windows version from master branch at https://binary-factory.kde.org/job/KMyMoney_Nightly_win64/. The same caveat applies. Also, if you reply to a bug by email, please trim everything you are replying to, as the entire message gets included as the next comment to the bug, so it just duplicates the previous comment.
(In reply to Jack from comment #8) I'm actually on bugs.kde.org but understood. Just tried 5.1.80-76ee0e-4df and looks like this is resolved. Additionally, it seems that what the big did was remove the investments from the report filter under accounts, and now I can see them. Which from memory was the behavior in prior versions before this bug became apparent. The only other call-out is that prior to this bug, investments were indented in the report, under the investment account, but now they are not. Not critical, but nice to have. Great work, I will continue with my daily driver, alternating with newer builds as needed.
Not sure if relevant, but I was testing something else and this bug appears to be back on kmymoney-master-1783-windows-cl-msvc2019-x86_64.7z
There is definitely something strange going on. If you just hit "Apply" on the Range tab of the original report, the data all seems to disappear. Saving the file with the new report and comparing the existing R000084 and copy of it from the XML file, I get: > budget="B000001" dataRangeStart="1" | dataRangeStart="nan" dataRangeEnd="10" | dataRangeEnd="nan" dataMajorTick="1" | dataMajorTick="0" (edited to get approx alignment with non-fixed-width font, and removing the id and name differences) Note that no budget shows up in either config, and both configs show All Dates and are are labelled 12/1/08 through 12/31/22.
Simpler test is to open Report000084, click Configure, click Apply, the data all disappears. Click Undo, the data does NOT reappear. Close and reopen the report, and the data is all present again. Among other problems, the display is apparently not always updated when the underlying data (either transaction or configuration data) is changed. What is even stranger is that simply clicking Apply only seems to lose data in some reports but not others (only tried within Net Worth reports so far.).
What appears to be the same bug was just brought up on https://discuss.kde.org/t/wierd-behavior-when-using-filters-in-reports/4624/1 and digging into the Configure dialog Accounts tab, the individual securities accounts (within the various investment accounts) are not listed. I need to dig into the source code, but I wonder if they are being somehow filtered out of the report, or just not even being considered for inclusion.