ENH: Enable ExcelWriter to construct in-memory sheets #10376
Conversation
bashtage
force-pushed
the
improve-excel
branch
2 times, most recently
from
96d6fe4 to
7197f8f
Compare
| @@ -249,10 +249,9 @@ Optional Dependencies | |||
| * `statsmodels <http://statsmodels.sourceforge.net/>`__ | |||
| * Needed for parts of :mod:`pandas.stats` | |||
| * `openpyxl <http://packages.python.org/openpyxl/>`__, `xlrd/xlwt <http://www.python-excel.org/>`__ | |||
| * openpyxl version 1.6.1 or higher, but lower than 2.0.0 | |||
There was a problem hiding this comment.
did you mean to take this out?
There was a problem hiding this comment.
Yes - both 1.x and 2.x are clearly supported and tested in excel.py and test_excel.py. I suppose the person who added the OpenPyXL 2.x path forgot to change this.
There was a problem hiding this comment.
hmm, IIRC there were some issues with >=2.0.0 and < 2.0.3 something like that (e.g. some style sheet issues. But ok, past that now, so ok.
jreback
added
Bug
Performance
|
|
||
| if sheet.nrows == 0: | ||
| return DataFrame() | ||
|
|
||
| if header is not None: |
There was a problem hiding this comment.
what about sheet.nrows==1 (and its the header)? do we construct an empty frame with the correct columns?
There was a problem hiding this comment.
Yes, I added a test for this case and it works correctly already.
There was a problem hiding this comment.
yep, saw that below. thanks.
| @@ -2184,6 +2184,38 @@ argument to ``to_excel`` and to ``ExcelWriter``. The built-in engines are: | |||
|
|
|||
| .. _io.clipboard: | |||
There was a problem hiding this comment.
move your doc to before this refernce (for the next section)
bashtage
force-pushed
the
improve-excel
branch
3 times, most recently
from
38b74a9 to
465ceb0
Compare
|
@jreback The vbench for openpyxl is very slow ~ around 300s on my computer. This is around 10x slower than the next slowest test (sql). Should I reduce the df size for this test? |
|
yes, a vbench should be < 1s if at all possible, so reduce the size on these. |
bashtage
force-pushed
the
improve-excel
branch
2 times, most recently
from
e64196c to
07fd4d6
Compare
|
I think it is finished once green - I snuck in one more Excel-related issue (testing xlwt on Py34). Openpyxl is very slow compared to xlsxwriter (5x slower) |
|
ahh, the yes, ping when green. |
|
@jreback ready |
| @@ -2182,6 +2184,40 @@ argument to ``to_excel`` and to ``ExcelWriter``. The built-in engines are: | |||
|
|
|||
| df.to_excel('path_to_file.xlsx', sheet_name='Sheet1') | |||
|
|
|||
There was a problem hiding this comment.
can you add a reference here (so we can link to it)
|
minor doc comment. ping after you have pushed for merging. |
|
@jreback Added those. |
Add support for StringIO/BytesIO to ExcelWriter Add vbench support for writing excel files Add support for serializing lists/dicts to strings Fix bug when reading blank excel sheets Added xlwt to Python 3.4 builds closes pandas-dev#8188 closes pandas-dev#7074 closes pandas-dev#6403 closes pandas-dev#7171 closes pandas-dev#6947
|
Rebased |
ENH: Enable ExcelWriter to construct in-memory sheets
|
@bashtage thanks for this! |
Add support for StringIO/BytesIO to ExcelWriter
Add vbench support for writing excel files
Add support for serializing lists/dicts to strings
Fix bug when reading blank excel sheets
closes #8188
closes #7074
closes #6403
closes #7171