Numbers 13 Bug Fixes (NOT)

I won't go into great detail unless you need more.

• Sorting of COUNTIFS I mentioned already. I think it is the most glaring of the bugs. I believe it started in version 10.
• The calculation engine gives incorrect results when doing repetitive calculations with small numbers. The results go haywire as compared to Excel and Matlab, which agree with each other. Numbers probably gave correct results prior to the new engine, which was new to version 10 (?). The new engine strayed away from the gold standard of IEEE 754, and not in a good way. This is actually a lot worse than the COUNTIF bug because it gives incorrect results to calculations. Correct mathematical calculations are the most important requirement for a spreadsheet app and it is failing that here.
• Erroneous value labels on charts. If a chart is displaying value labels and the first or last Y data point in the series is a blank cell, it will put a value label on the X axis for that non-existent point as if the point was a zero. This is most annoying when using "series name" value labels. Other blank cells do not get value labels, just the first and last ones. I don't know when it started but it was at least a year ago. Plotting a second series with only a single data point and with Series Name value labels is a way to highlight and label a specific point on a chart but it is useless with this bug.
• The result of SUMIFS, when dates are the conditions, will change if the file region is changed. I only tested this with Canada vs US because that was the original poster's problem. If the spreadsheet is designed in the US then the region changed to Canada, the results change. This bug was in version 11.2, maybe earlier.
• TEXTJOIN when joining a range that has empty cells will not ignore the empty cells, except for the first cell in the range. When using a separator, it results in repeated separators in the list of words (such as One,,Three,Four). This bug was present when TEXTJOIN was released.
• CONCAT and TEXTJOIN can both be used to create a "blank" cell, one where ISBLANK returns TRUE even though the cell obviously has a formula in it. Whether this is a bug or an undocumented feature is debatable.
• There are cases where a chart will not create a trendline even when one exists. The trendline equation simply says Y=0. The correct trendline can be calculated in the table and plotted as a second series as a workaround. Excel calculates the trendline correctly. This was first noticed in January 2023.
• WORKDAY(date, work-days, exclude-dates) when there are "exclude-dates" gives erroneous results. It will skip over perfectly valid workdays if "date" is an exclude date and "work-days" is +/-1. For instance if the exclude date is Tuesday, it will say the previous workday was Friday, not Monday. The problem may go deeper than that, this is the only test I did. Excel gets it right. I did not write down when this bug was first noticed but my spreadsheet is using dates in May 2021. I know for sure it has been around since before April 2022.
• Borders not displaying or printing correctly. These are two related bugs. To see it, make all cells and all borders black. De-select the table and it should be all black, but it is not. If the border are less than 1pt it will have white borders around cells (bug 1). Change it so there are no borders in the entire table. Now it will be black on the screen but if you print it it will have faint white borders around all cells (bug 2). The only way for it to be black on the screen and black when printed is for the borders to be 1 pt or larger.
• MATCH will do what appears to be a bottom to top search although the documentation says it should be doing a top to bottom search (and Excel does it top to bottom). If A1 and A7 have the letter "a", MATCH("a", A1:A7,1) incorrectly gives the bottom result (7). MATCH("a",A1:A7,-1) and MATCH("a",A1:A7,0) give the top result (1). All should have given the top result.

There have been other bugs mentioned here but I got tired of writing them up, reporting them, creating spreadsheets to demonstrate them, and being disappointed with every new release that didn't fix them. Looking at file dates of my "bug" files, the last time a bug I was tracking was fixed was a year ago, April 2022.