When a CSV with time fields formatted as strings is imported, the resulting parsed time value isn’t precise enough, which then results in rounding errors in the displayed time. It seems that there are two bugs happening here: First, the import is imprecise, second, the timestamp display has rounding issues.
A timestamp such as “8:00:00 am” is parsed into the time/float format as 0.333333, which then displays as 7:59:60 AM. In case it’s relevant, entering 0.333333 and manually formatting the cell as a time value results in the same rounding error.
Notably, typing out “8:00 am” into a spreadsheet directly results in the more precise parsed value 0.333333333333333, which displays correctly as 8:00 AM. This only manifests on values with repeating decimals.
Interestingly, when using Data > Get Data > From Local TXT/CSV, the time value is parsed more precisely, as 0.333333333333333, which in turn means it displays correctly. However, that parser has issues with newlines in strings, so it is not a useful workaround, at least for my use case. Perhaps that is worth a second bug report.
As far as I can tell, this is a bug, not malformed input or other user error. Additionally, Excel (for web) parses all of these values correctly. Timestamps in CSVs display correctly (due to web Excel, I am not able to check the true converted float here, but entering the time manually gets 0.333333333333333, like ONLYOFFICE), and 0.333333 displays as 8:00 AM as expected.
I am a new user and cannot upload attachments yet, so here is some example CSV data that reproduces the issue.
“Time”,“Info”
“8:00:00 am”, “This is a test CSV”
“9:00 am”, “to demonstrate the issue”
“10:00 am”, “with timestamps”
“11:00 am”, “when importing a CSV”
OS version: MacOS 26.2, and Windows 11 23H2.
App version: 9.2.1
Downloaded from: ONLYOFFICE website
Additional information: Present on 9.2 as well, older versions untested.