[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 gallon.fi...@gmail.com changed: What|Removed |Added Resolution|--- |FIXED Status|NEW |RESOLVED --- Comment #7 from gallon.fi...@gmail.com --- Fixed via r1850212. I rework both test and implementation but thanks to Axel Howind for the report and the primary idea. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 Axel Howind changed: What|Removed |Added OS|Mac OS X 10.1 |All Hardware|PC |All -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 Axel Howind changed: What|Removed |Added Attachment #36313|0 |1 is obsolete|| Attachment #36315|0 |1 is obsolete|| --- Comment #5 from Axel Howind --- Created attachment 36319 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36319=edit Patch with bugfix and testcase Revised patch, now all unit tests pass. I have also added a new test case for this. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 --- Comment #4 from Axel Howind --- Sorry, it seems this patch introduces a regression that I didn't see because Junit was broken on my machine (I opened Bug 62996 for that problem). Now that Junit is up and running again, I will look into the regression this has created and create a updated patch. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 --- Comment #3 from Axel Howind --- Created attachment 36315 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36315=edit Patch that fixes bug 62993 Patch to fix Bug 62993 `XSSFEvaluationSheet` uses an internal cache to speed up evaluations. It also contains an integer field `_lastDefinedRow` to avoid calling `getLastRowNum()` for the spreadsheet which is costly because it looks up the last entry in a sorted map which should be O(log(n)) for the implementation used (with n being the number of rows). The problem is, that `_lastDefinedRow` is evaluated only once when the `XSSFEvaluationSheet` is created and not updated when rows are added later. This patch removes `XSSFEvaluationSheet._lastDefinedRow` and instead always uses `_xs.getLastRowNum()`. Runtime performance should be unaffected or even better because I moved the optimization (caching of the last row number) up to `XSSFSheet`, so that other code will also benefit. Please review and comment. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 --- Comment #2 from Axel Howind --- It turns out that this probably not only affects SUMIF but possibly all functions that operate on ranges. Bug is triggered when rows are added after the first formula evaluation is done. I have prepared a patch and will include a more detailed description there. Please note that this is almost certainly a regression from POI 3.x since the Ruby code we use produced correct results before upgrading, and the changes during upgrade were rather cosmetically (exchanging ints with enums in some places). -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
[Bug 62993] SUMIF returns wrong result (POI 4.0.x)
https://bz.apache.org/bugzilla/show_bug.cgi?id=62993 --- Comment #1 from Axel Howind --- Created attachment 36313 --> https://bz.apache.org/bugzilla/attachment.cgi?id=36313=edit JUnit test for this The attachment contains a JUnit test case for this bug. I am unsure where this should end up in the POI source tree though. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@poi.apache.org For additional commands, e-mail: dev-h...@poi.apache.org
