DEPR: ExcelFile.parse

[rhshadrach]

It seems to me we should either fix ExcelFile.parse or deprecate it entirely, and I lean toward the latter. pandas originally started out with just ExcelFile but now has the top-level read_excel. The signatures started the same, but now read_excel has gained and modified parameters that have not been added/changed in ExcelFile.parse. For example:

• ExcelFile.parse lacks a dtype parameter
• ExcelFile.parse has a **kwds argument that is passed on to pandas internals with no documentation on what can be included. Invalid arguments are just ignored (e.g. BUG: xl.parse index_col ignoring skiprows #50953)

It appears to me that pd.ExcelFile(...).parse(...) offers no advantage over pd.read_excel(pd.ExcelFile(...)), and so rather than fixing parse we can deprecate it and make it internal.

Edit: I no longer think deprecating ExcelFile entirely as mentioned below is a good option. See #58247 (comment).

Another option is to deprecate ExcelFile entirely. The one thing ExcelFile still provides that isn't available elsewhere is to get the underlying book or sheet_names without reading the entire file.

df = pd.DataFrame(np.zeros((100, 100)))
with pd.ExcelWriter("test.xlsx") as writer:
    for e in range(10):
        df.to_excel(writer, sheet_name=str(e))

%timeit pd.ExcelFile("test.xlsx").sheet_names
# 14.1 ms ± 76 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)
%timeit pd.read_excel("test.xlsx", sheet_name=None)
# 411 ms ± 2.07 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

One can somewhat work around this by using nrows, but it's clunky.

%timeit pd.read_excel("test.xlsx", sheet_name=None, nrows=0).keys()
# 57.3 ms ± 257 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

[Aloqeely]

Yes I agree we should deprecate ExcelFile.parse, no reason to fix it.

And I'm not against deprecating ExcelFile either.

[rmhowe425]

+1 for deprecating ExcelFile and ExcelFile.parse

[samukweku]

To facilitate working with multiple sheets from the same file, the ExcelFile class can be used to wrap the file and can be passed into read_excel There will be a performance benefit for reading multiple sheets as the file is read into memory only once.

Deprecating Excel file.parse means this feature is lost. What would be the alternative @rhshadrach ?

[asishm]

@samukweku That is not lost if only pd.ExcelFile.parse is deprecated while pd.ExcelFile remains available (which is what I would favor as well).

[rhshadrach]

Deprecating Excel file.parse means this feature is lost. What would be the alternative @rhshadrach ?

pd.read_excel

[samukweku]

@asishm kindly explain how it is not lost? @rhshadrach wont the pd.read_excel option be less performant for multiple sheets( since with read_excel you read it more than once)?

[rmhowe425]

@samukweku the read_excel() function does allow you to read in multiple sheets at once by explicitly specifying the worksheets to be read in as values for the sheet_name parameter. Or you can pass None as a value to read in all sheets.

Albeit, when reading in multiple worksheets you are returned a dict of DataFrames.

[rhshadrach]

@rhshadrach wont the pd.read_excel option be less performant for multiple sheets( since with read_excel you read it more than once)?

You can pass an ExcelFile instance to read_excel. It directly calls ExcelFile.parse internally.

[asishm]

As noted in the user guide, the usecase below can't be achieved without a re-read with the current api of pd.read_excel(fp-like)

The primary use-case for an ExcelFile is parsing multiple sheets with different parameters:

data = {}
# For when Sheet1's format differs from Sheet2
with pd.ExcelFile("path_to_file.xls") as xls:
    data["Sheet1"] = pd.read_excel(xls, "Sheet1", index_col=None, na_values=["NA"])
    data["Sheet2"] = pd.read_excel(xls, "Sheet2", index_col=1)

[rhshadrach]

Thanks @asishm - that has me convinced that ExcelFile itself should stay.

[GAuravY19]

what is supposed to be done in this issue ??

Can you please clarify it properly.