BUG: inconsistent and undocumented option "converters" to read_excel #8548
Conversation
|
usually start off a PR with tests first. |
|
I added a test for this functionality. |
|
ok, this looks fine
|
jreback
added
Docs
IO Excel
|
I added some docs there, not sure it's enough, open for feedback. |
| @@ -1992,6 +1992,31 @@ indices to be parsed. | |||
|
|
|||
| read_excel('path_to_file.xls', 'Sheet1', parse_cols=[0, 2, 3]) | |||
|
|
|||
| .. versionadded:: 0.15 | |||
There was a problem hiding this comment.
take this line out (it has existed previously, just not documented)
|
Hi, what's the current status on this? I think I implemented all you requested, do you think it's ready and can be merged? |
| It is possible to manipulate the contents of single Excel cells while reading | ||
| via the `converters` option. `converters` is a dictionary of functions: the keys | ||
| are the names or indices of columns to be transformed, the values are functions | ||
| that take one input argument, the Excel cell content, and return the transformed |
There was a problem hiding this comment.
take out the sentence starting with 'Lambda function'; its uncessary. Can you make this a bit more concise?
|
ok, pls edit the docs a bit, rebase and squash. |
iosonofabio
force-pushed
the
excel_dtype
branch
from
087fff5 to
d0597b8
Compare
|
Ok, edited docs, squashed, rebased, pushed (d0597b8). |
|
|
||
| It is possible to transform the contents of Excel cells via the `converters` | ||
| option. It accepts a dictionary of functions: the keys are the names or | ||
| indices of columns to be transformed, the values are functions that take one |
There was a problem hiding this comment.
instead of this, I think just expand the doc-string (just a bit, maybe 1 line) for TextReader and Excel (to make them consistent). This is also true more generally for csv-type reading, so don't want it specifically here.
There was a problem hiding this comment.
I am sorry, could you please reformulate? I can't understand.
- What do you want to delete from the docs? I thought you wanted to have the ..note, or shall I delete the whole thing again?
- Which line do you want to add to those docstrings? Could you please write here the line you want?
- TextReader? or TextParser?
- What is true more generally?
Thanks.
edit: I tried to implement what I understood of your suggestion, please check and let me know.
| Dict of functions for converting values in certain columns. Keys can | ||
| either be integers or column labels, values are functions that take one | ||
| input argument, the cell (not column) content, and return the | ||
| transformed content. | ||
| encoding : string, default None |
|
looks ok to me |
BUG: inconsistent and undocumented option "converters" to read_excel
|
Thanks! |
|
Thanks! I'm happy since it's my first PR and I made a bit of a mess. I'll try to work on the full dtype integration, using the C parser. |
Issue #8212 (first part): pandas.read_excel accepts an optional argument "converters" (which is passed down to PythonParser) to convert single cells in columns with a conversion function. I documented this feature and added a try/except block to make it work in case some cells contain NaNs.
What's still missing is the full "dtype" argument, a la read_csv. That patch is somewhat orthogonal because it only works with the C parser, so I plan to implement it in a second step.