BUG: fix precision to specify decimal places #13897
Conversation
closes GHI13257 Use 'f' as a string conversion type to specify the number of decimals to represent. The 'g' format will instead fall back to the 'e' format causing the unexpected behavior seen in GHI13257. See: https://docs.python.org/3.5/library/stdtypes.html#old-string-formatting
| ctx = df.style.format({"a": "{:0.1f}", "b": "{0:.2%}"}, | ||
| subset=pd.IndexSlice[0, :])._translate() | ||
| expected = '0.1' | ||
| precise_11 = '1.123400' # with the default precision |
There was a problem hiding this comment.
pls also test changing the precision from default.
| Also, add tests with |
Current coverage is 85.28% (diff: 100%)@@ master #13897 diff @@ ========================================== Files 139 139 Lines 50057 50057 Methods 0 0 Messages 0 0 Branches 0 0 ========================================== - Hits 42692 42691 -1 - Misses 7365 7366 +1 Partials 0 0
|
| @drewrey pls add a whatsnew note (bug fix section) |
| Changing it to In [1]: "{:f}".format(1) Out[1]: '1.000000'and instead have that formatted as Could you add a test with an integer and a float column like |
| Looking further, it seems like we have a bug in master. m = pd.DataFrame({"a": [1], "b": [1.]})
Ideally |
| can you rebase / update? |
| can you add a whatnew note in 0.19.2 |
| closing, but if you'd like to update, pls do so. |
5 participants
git diff upstream/master | flake8 --diffUse 'f' as a string conversion type to specify the number of decimals to
represent. The 'g' format will instead fall back to the 'e' format
causing the unexpected behavior seen in GHI13257.
See: old-string-formatting
Replicated code from the issue, with output:
