I have a very simple for loop which was working last week but when I run it this week on exactly the same data it is not working... when I run the code outside of the loop it works without an error. I don't understand...

The code is: '''

survey = pd.read_csv(folderin + 'survey.csv')
CSsurvey = pd.read_csv(folderin + 'CS_survey.csv')

sports = ['Value MTB related (e.g. inner tubes, water bottles etc)',
'Value Running','Value Roaming and other outdoor related (e.g. climbing, kayaking)',
'Value Outdoor sports event related (e.g.race)']

sport = []
for p in sports:
    item = survey[p].sum()
    CSitem = CSsurvey[p].sum()
    total = item + CSitem
    sport.append(total)'''

I get the error:

If I run the code for each 'p' outside of the loop I don't get an error

I have a very simple for loop which was working last week but when I run it this week on exactly the same data it is not working... when I run the code outside of the loop it works without an error. I don't understand...

The code is: '''

survey = pd.read_csv(folderin + 'survey.csv')
CSsurvey = pd.read_csv(folderin + 'CS_survey.csv')

sports = ['Value MTB related (e.g. inner tubes, water bottles etc)',
'Value Running','Value Roaming and other outdoor related (e.g. climbing, kayaking)',
'Value Outdoor sports event related (e.g.race)']

sport = []
for p in sports:
    item = survey[p].sum()
    CSitem = CSsurvey[p].sum()
    total = item + CSitem
    sport.append(total)'''

I get the error:

If I run the code for each 'p' outside of the loop I don't get an error

• pandas
• loops
• keyerror

• what does the data look like? have you confirmed that the column headers you iterate through exist in each data frame? if you cannot be sure at run time you can add logic to check if that column header exists before attempting to index – UnsanitizedInput Commented Mar 24 at 15:37
• I have copied and pasted each of the strings in sports = [] and run the code with these. No issue at all. – Heather Kay Commented Mar 24 at 15:43
• Seems like the solution is to close spyder and reopen it... lets see if it still works tomorrow.... – Heather Kay Commented Mar 24 at 16:16
• 1 sport = pd.concat([survey, CSurvey]).sum(axis=0)[sports].to_list() – Scott Boston Commented Mar 24 at 20:27
• Thanks @ScottBoston I knew there was an easier way for me to do this but hadn't realised you could concat with a specific set of columns. Awesome. – Heather Kay Commented Mar 25 at 8:44

1 Answer 1

IIUC then you can rewrite the above for loop using pd.concat and then taking the sum over all columns and using column filtering to return on the sports columns finally add to_list to convert those sums to a list.

Try:

sport = pd.concat([survey, CSurvey]).sum(axis=0)[sports].to_list()