使用python脚本将文件从.xls转换为.csv时，如何使数字不显示为指数？

Question

I am trying to insert excel data into a table (postgres) using a python script. I am running into an issue though where some of the larger numbers get inserted as exponentials. I realized this is happening when I am converting the file from .xls to .csv (I never open the .xls files because I realize that excel does some funky stuff where it'll save larger numbers into exponential form)

Is there an easy way to ensure the numbers don't get displayed as exponentials?

ie 812492400097 is being displayed as 8.12E+11

Here is the convert to .csv script:

import xlrd    
import unicodecsv    
import sys    
import os    
import datetime


def csv_from_excel(xlsfile, csvfile):  
    wb = xlrd.open_workbook(xlsfile)  
    sh = wb.sheet_by_index(0)  
    outputfile = open(csvfile, 'wb')  
    wr = unicodecsv.writer(outputfile, quoting=unicodecsv.QUOTE_ALL)

    for rownum in xrange(sh.nrows):
        wr.writerow(sh.row_values(rownum))

    outputfile.close()

def log(s):
    print str(datetime.datetime.now()) + ": " + s


#main
if len(sys.argv) < 2:   
    print "Missing parameters: input xls file"   
    sys.exit()

sourcefile = sys.argv[1]

destfile = sourcefile.split('.')[0] + '.csv'

log("processing " + sourcefile)

csv_from_excel(sourcefile, destfile)

Also wondering if perhaps instead of ensuring the .csv doesn't turn numbers to exponentials, to turn exponentials to numbers when inserting into the postgres table?

Answer 1

The xlrd module treats all numbers from Excel as floats, because Excel calculates all numbers as floats:

https://secure.simplistix.co.uk/svn/xlrd/trunk/xlrd/doc/xlrd.html?p=4966#sheet.Cell-class

• Type symbol [Type number] Python value
• XL_CELL_EMPTY [0] empty string u''
• XL_CELL_TEXT [1] a Unicode string
• XL_CELL_NUMBER [2] float
• XL_CELL_DATE [3] float
• XL_CELL_BOOLEAN [4] int; 1 means TRUE, 0 means FALSE
• XL_CELL_ERROR [5] int representing internal Excel codes; for a text representation, refer to the supplied dictionary error_text_from_code
• XL_CELL_BLANK [6] empty string u''. Note: this type will appear only when open_workbook(..., formatting_info=True) is used.

Your solution, presumably, needs to define the formatting used by unicodecsv to write floats.

A previous question ( How can I prevent csv.DictWriter() or writerow() rounding my floats? ) indicates that the csv module used to use float.__str__ rather than float.__repr__ , which caused rounding. unicodecsv might still use float.__str__ .