Need Pivot Table for Columns that Have Comma Separated Values

Question

So, I found a couple other solutions to similar questions I have, but not quite exact. I am interpreting survey results in Excel where the survey tool (Qualtrics) has placed responses from multiple select questions ("select all that apply") in a single cell, comma separated. I need counts of the individual responses calculated in a Pivot Table, where I will also take totals from some respondent demographics I'm going to add to the response spreadsheet. What I am trying to do is very similar to this:

Split comma separated entries to new rows

However my sheet will have multiple columns with comma separated responses, like this:

Existing data

....and I need it in column format so I can Pivot and count. As I mentioned I am going to add in some attribute data (HR data, and that's why I cannot import it into Qualtrics - can't send outside the company), so I'll still need to pair up the person with the response. However, if there is a script or command that can run and split this out for me, I'm not sure how it would handle differing numbers of response from column to column to create the needed rows (Like Bill and Karen in the example). Would I need to have the column with the longest csv cell first and so on? However, it would look something like this:

Desired Data

Is there something I can do to accomplish this?

score 5 · Accepted Answer · answered Jan 16 '19 at 00:19

your desired layout is actually not fit for proper pivot tables, either. You need a really flat table structure, so you can filter on likes without simultaneously hiding dislikes.

You can easily transform your data with Power Query. Load the data into the Power Query Editor, then split each question column by the delimiter ", " (comma followed by space). This will split each answer into its own column, with the question in the header appended by .1, .2 etc.

Then select the name column and click "Unpivot other columns". The question headers will now be in the attribute column. Split that attribute column by the delimiter "." (dot) and delete the column with the split off numbers.

Finally, rename the columns to Question and Answer.

Here is the M code that is generated when doing that.

let
    Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Name", type text}, {"like", type text}, {"dislike", type text}}),
    #"Split Column by Delimiter" = Table.SplitColumn(#"Changed Type", "dislike", Splitter.SplitTextByDelimiter(", ", QuoteStyle.Csv), {"dislike.1", "dislike.2", "dislike.3", "dislike.4", "dislike.5"}),
    #"Changed Type1" = Table.TransformColumnTypes(#"Split Column by Delimiter",{{"dislike.1", type text}, {"dislike.2", type text}, {"dislike.3", type text}, {"dislike.4", type text}, {"dislike.5", type text}}),
    #"Split Column by Delimiter1" = Table.SplitColumn(#"Changed Type1", "like", Splitter.SplitTextByDelimiter(", ", QuoteStyle.Csv), {"like.1", "like.2", "like.3", "like.4"}),
    #"Changed Type2" = Table.TransformColumnTypes(#"Split Column by Delimiter1",{{"like.1", type text}, {"like.2", type text}, {"like.3", type text}, {"like.4", type text}}),
    #"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type2", {"Name"}, "Attribute", "Value"),
    #"Split Column by Delimiter2" = Table.SplitColumn(#"Unpivoted Other Columns", "Attribute", Splitter.SplitTextByDelimiter(".", QuoteStyle.Csv), {"Attribute.1", "Attribute.2"}),
    #"Changed Type3" = Table.TransformColumnTypes(#"Split Column by Delimiter2",{{"Attribute.1", type text}, {"Attribute.2", Int64.Type}}),
    #"Removed Columns" = Table.RemoveColumns(#"Changed Type3",{"Attribute.2"}),
    #"Renamed Columns" = Table.RenameColumns(#"Removed Columns",{{"Attribute.1", "question"}, {"Value", "Answer"}})
in
    #"Renamed Columns"

The resulting table looks like this:

If your original data receives more rows, just refresh the query.

Now you have a table that can be used in pivots without compromising any data.

Using the resulting table above, how would I count the number of "dislikes" in a pivot table "value" without overcounting the "Karen" rows? — tquill, Aug 30 '23 at 20:41

score 0 · Answer 2 · answered Jan 16 '19 at 18:16

Not sure exactly what do you want to achieve (may be too many things at once), at least for the first part of your question: Count of response per category you can do this:

For each response column, create a count column, for example:

Name, Response_column_a, count_a, ...

The formula for count: count of comma + 1

In excel it can be acchieved with the following formula:

You now have: 1) count of responses 2) can generate total count of responses per category, for example count of Response A for Karen.

Need Pivot Table for Columns that Have Comma Separated Values

2 Answers2