如何使用 Python 修改 SPSS 输出文件？

Question

I'm making custom tables in SPSS, but when the cell values (percentages) are rounded to 1 decimal, they sometimes add up to 99,9 or 100,1 in stead of 100,0. My boss asked my to have everything neatly add up to 100. This means slightly changing some values in the output tables.

I wrote some code to retrieve cell values from tables, which works fine, but I cannot find any method or class that allows me to change cells in already generated output. I've tried things like :

Table[(rij,6)] = CellText.Number(11)

and

SpssDataCells[(rij,6)] = CellText.Number(11)

but it keeps giving me "AttributeError: 'SpssClient.SpssTextItem' object has no attribute 'DataCellArray'"

How do I succesfully change cell values of output tables in SPSS?

My code so far:

import SpssClient, spss

# Python verbinden met SPSS.
SpssClient.StartClient()
OutputDoc = SpssClient.GetDesignatedOutputDoc()
OutputItemList = OutputDoc.GetOutputItems()

# Laatste tabel pakken.
lastTab = OutputItemList.Size()-2
OutputItem = OutputItemList.GetItemAt(lastTab)
Table = OutputItem.GetSpecificType()
SpssDataCells = Table.DataCellArray()

# For loop. Voor iedere rij testen of de afgeronde waarden optellen tot 100.
# Specifieke getallen pakken.
rij=0
try:
while (rij<20):
   b14 = float(SpssDataCells.GetUnformattedValueAt(rij,0))
   z14 = float(SpssDataCells.GetUnformattedValueAt(rij,1))
   zz14 = float(SpssDataCells.GetUnformattedValueAt(rij,2))
   b15 = float(SpssDataCells.GetUnformattedValueAt(rij,4))
   z15 = float(SpssDataCells.GetUnformattedValueAt(rij,5))
   zz15 = float(SpssDataCells.GetUnformattedValueAt(rij,6))
   print [b14,z14,zz14,b15,z15,zz15]
   rij=rij+1
except:
  print 'Einde tabel'

Answer 1

The SetValueAt method is what you require to change the value of a cell in a table.

PS. I think your boss should focus on more important things than to spend billable time on having percentages add up neatly to 100% (due to rounding). Also ensure you are using as many decimal point precision as possible in your calculations so to minimize this "discrepancy".

Update:

Just to give an example of what you can do with manipulation like this (beyond fixing round errors):

The table above shows the Share of Voice (SoV) of a Respiratory drug brand (R3) and it's rank among all brands (first two columns of data) and SoV & Rank also within it's same class of brands only (third and forth column). This is compared against previous month (July 15) and if the rank has increased then it is highlighted in green and an upward facing arrow is added and if declined in rank then highlighted in red and downward red facing arrow added. Just adds a little, color and visualization to what otherwise can be dull tables.