具有python字符串索引的XML到CSV的XML必须是整数

Question

I'm trying to convert an xml to csv but when i'm running the python script i'm getting this error:

TypeError: string indices must be integers

The XML structure ( it's bigger but its always the same ):

<?xml version='1.0' encoding='UTF-8'?>
<import>
  <products>
    <product>
      <attribute>
        <code>Something</code>
        <value>xxx</value>
      </attribute>
      <attribute>
        <code>Something2</code>
        <value>xxx</value>
      </attribute>
      <attribute>
        <code>Something3</code>
        <value>xxx</value>
      </attribute>
      <attribute>
        <code>Something4</code>
        <value>xxx</value>
      </attribute>
    </product>
  </products>
</import>

The python file:

#!/usr/bin/python
# -*- coding: utf-8 -*-
import csv, xmltodict
from collections import OrderedDict
class Test:
    def PSXML(self):
        FilePS = open('test.csv', 'w')
        csvwriter = csv.writer(FilePS)
        header = ['Something1','Something2','Something3','Something4']
        csvwriter.writerow(header)
        with open('test.xml') as fd:
            PSdata = []
            obj = xmltodict.parse(fd.read())
            obj = obj['import']['products']
            root_elements = obj['product'] if type(obj) == OrderedDict else [obj['product']]
            for element in root_elements:
                Something1 = element['attribute'][1]['value']
                PSdata.append(Something1)
                Something2 = element['attribute'][2]['value']
                PSdata.append(Something2)
                Something3 = element['attribute'][3]['value']
                PSdata.append(Something3)
                Something4 = element['attribute'][4]['value']
                PSdata.append(Something4)
                csvwriter.writerow(PSdata)
        FilePS.close()
TryIT = Test()
TryIT.PSXML()

This code already worked with another XML structure ( a more logic one ) but on this one is crashing with that TypeError: string indices must be integers error.

Any one have an idea why is that ?

Answer 1

The problem here is, that in your example you just got one product. So as Elis said, in the loop element is just the string attribute .

Actually you already tried to cover the possibility of just one product but you made a mistake there. You have to put it in a list, if it IS an OrderedDict:

root_elements = obj['product'] if type(obj) != OrderedDict else [obj['product']]

The code works well if there are several products except a couple of small mistakes.

You have to put the PSdata initialization inside the loop where you iterate over your products, because otherwise, with every product, 4 new columns get attached and the values of the former product written in the same line.

So you might want to check if there is only one product in your root_elements and handle this case separately.

Furthermore don't name your variables uppercase.

Another point: lists are zero indexed in Python, so to get the 4 value you should go with:

for element in root_elements:
    psdata = []
    something1 = element['attribute'][0]['value']
    psdata.append(something1)
    something2 = element['attribute'][1]['value']
    psdata.append(something2)
    something3 = element['attribute'][2]['value']
    psdata.append(something3)
    something4 = element['attribute'][3]['value']
    csvwriter.writerow(psdata)

Or shorter with a list comprehension:

for element in root_elements:
    csvwriter.writerow([element['attribute'][i]['value'] for i in range(4)])

So here is an updated version of your script, which follows most of pep8:

#!/usr/bin/python
# -*- coding: utf-8 -*-
import csv
import xmltodict
from collections import OrderedDict


class Test:

    def psxml(self):
        with open('test.csv', 'w') as file_ps:
            csvwriter = csv.writer(file_ps)
            header = ['Something1', 'Something2', 'Something3', 'Something4']
            csvwriter.writerow(header)
            with open('test.xml') as fd:
                obj = xmltodict.parse(fd.read())
                obj = obj['import']['products']
                root_elements = obj['product'] if type(obj) != OrderedDict else [obj['product']]
                for element in root_elements:
                    csvwriter.writerow([element['attribute'][i]['value'] for i in range(4)])

try_it = Test()
try_it.psxml()

Answer 2

Try:

for element in root_elements:
    print element, type(element)

will print:

attribute <type 'unicode'>

You may expect a dict but it is a string.