Save large pandas dataframe to excel

Question

I'm generating a large dataframe (1.5 GB when saved in CSV format) and need to store it an worksheet of an Excel file along with a second (much smaller) dataframe which is saved in a separate worksheet.

print('Reading temporaty files for variable {}:'.format(Var))
print(' Reading stations')
s=pd.read_csv(StatFile,sep=':',dtype={'ID': 'str'},encoding='utf-8')
print(' Reading data')
d=pd.read_csv(DataFile,sep=':',dtype='str',encoding='utf-8').transpose()
d.columns = d.iloc[0]
d=d[1:].astype('float')
d.reindex_axis(sorted(d.columns), axis=1)
print('Writing out Excel file for variable {}'.format(Var))
writer = pd.ExcelWriter(Path + Var + '.xlsx', engine='xlsxwriter')
d.to_excel(writer, sheet_name='Data')
OutStatCol=['ID','Name','Longitude','Latitude','GRS','OriginalVariable','VariableUnits','URL','JsonNode']
s.to_excel(writer, columns=OutStatCol, index=False, sheet_name='Stations')
writer.save()

My code works fine for smaller dataframes, but with the large ones I get the following error:

Traceback (most recent call last):
  File "./Test2.py", line 29, in <module>
    writer.save()
  File "/home/user/miniconda2/lib/python2.7/site-packages/pandas/io/excel.py", line 1413, in save
    return self.book.close()
  File "/home/user/miniconda2/lib/python2.7/site-packages/xlsxwriter/workbook.py", line 297, in close
    self._store_workbook()
  File "/home/user/miniconda2/lib/python2.7/site-packages/xlsxwriter/workbook.py", line 624, in _store_workbook
    xlsx_file.write(os_filename, xml_filename)
  File "/home/user/miniconda2/lib/python2.7/zipfile.py", line 1148, in write
    self._writecheck(zinfo)
  File "/home/user/miniconda2/lib/python2.7/zipfile.py", line 1114, in _writecheck
    " would require ZIP64 extensions")
zipfile.LargeZipFile: Filesize would require ZIP64 extensions

Is there any way I can specify something like allowZip64=True in the ExcelWriter declaration or in the to_excel() method?

Thanks!

Answer 1

This took some source code digging, but...

print('Reading temporaty files for variable {}:'.format(Var))
print(' Reading stations')
s=pd.read_csv(StatFile,sep=':',dtype={'ID': 'str'},encoding='utf-8')
print(' Reading data')
d=pd.read_csv(DataFile,sep=':',dtype='str',encoding='utf-8').transpose()
d.columns = d.iloc[0]
d=d[1:].astype('float')
d.reindex_axis(sorted(d.columns), axis=1)
print('Writing out Excel file for variable {}'.format(Var))
writer = pd.ExcelWriter(Path + Var + '.xlsx', engine='xlsxwriter')

#THIS
writer.book.use_zip64()

d.to_excel(writer, sheet_name='Data')
OutStatCol=['ID','Name','Longitude','Latitude','GRS','OriginalVariable','VariableUnits','URL','JsonNode']
s.to_excel(writer, columns=OutStatCol, index=False, sheet_name='Stations')
writer.save()

should work

figuring out that the writer didn't inherit from workbook took me longer than it should have. writer.book is directly a workbook instance... d'oh