Split Large Text Field

[silverdale9999]

Not so much of a defined question, more of a "does anyone have any ideas I could try" kind of thing.

One of our business processes utilises a large text field within a ticketing system. (Rather than the more obvious use of actual system fields.. but anyway). I can't change this process - tried, but it's embedded across a lot of teams and it's a no go area. So to use one of my "favourite" corporate expressions.. "we are where we are".

This text field has around 20 different "fields" within it. Each one denoted by a number the start.. eg:

20_Who is the approver for this work

22_Who is implementing this work

They all follow this format, so the actual splitting of the columns is ok-ish - I've used this in my PowerQuery:

 = Table.SplitColumn(#"Removed Blank Rows", "Notes", 
 Splitter.SplitTextByAnyDelimiter({"1_","2_","3_","4_","5_","6_" etc

This may not be the most elegant, but it works and gets me a bunch of columns that can be renamed, and text that can be cleaned up.

The issue I'm having, is when new "fields" are added to the wall of text. If they're added at the end, fine. I just need to amend my query and add to the number of columns to split out.

When they are added in the middle though, it causes issues as what was Column B, now becomes Column C, and the new data is added to Column B, giving me a mix of different data in the same column.

This data is loaded in it's entirety again every day from the source system via a spreadsheet loaded onto sharepoint.

One thought I had was to split the query.. so load data up to a certain date with 'Query A'. Then, after the notes field has changed, use 'Query B' that will follow slightly different rules for the split, then append B to A once done.

Not sure this is the best approach though, so if anyone has any other suggestions (apart from get a new job) I'd be really appreciative.

If there is a better approach than using the split columns I'm happy to try it, not set on any particular way.

Thanks!!

[mahoneypat]

Can you provide a few rows of example or mock data, and what your desired output would look like?  I'm sure there is an easier way but your scenario is a little unclear.  For example, why not just split by "-" instead of the list "1_", "2_", etc.?

If this works for you, please mark it as the solution.  Kudos are appreciated too.  Please let me know if not.

Regards,

Pat

[silverdale9999]

Thanks for the reply!

I originally was using the underscore to split, but then it became clear that some of the information contained within the fields also contains an underscore so it was causing quite a lot of "false" splits.

Some more info:

Example text:

4_ Describe the risks
<None>
5 _ Will this cause an outage
<Y>
6_What is the impact
<No Impact>
8_ Can this be regressed?
<Y>
20_Who is the approver for this work
<Joe Smith>
22_Who is implementing this work
<Joe Smith>

This should split to:

4_ Describe the risks	5 _ Will this cause an outage	6_What is the impact	8_ Can this be regressed?	20_Who is the approver for this work	22_Who is implementing this work
None	Y	No Impact	Y	Joe Smith	Joe Smith

This works fine, in the current format.

The issue arises when the format of the notes field is changed.  EG, adding question 7 in bold below:

4_ Describe the risks
<None>
5 _ Will this cause an outage
<Y>
6_What is the impact
<No Impact>
7_External Impact?
<N>
8_ Can this be regressed?
<Y>
20_Who is the approver for this work
<Joe Smith>
22_Who is implementing this work
<Joe Smith>

This results in the following.. The original format loads fine, but the new format does not - data is inserted into the wrong column, and data is missed off the end.

4_ Describe the risks	5 _ Will this cause an outage	6_What is the impact	8_ Can this be regressed?	20_Who is the approver for this work	22_Who is implementing this work
None	Y	No Impact	Y	Joe Smith	Joe Smith
None	Y	No Impact	N	Y	Joe Smithv

I need to find a way to keep the text with it's question.. but I can't figure out where to start!

Thanks

[v-zhenbw-msft]

Hi @silverdale9999 ,

We can use the following steps in Power Query Editor to meet your requirement.

1. Add a conditional column, if it doesn’t contain “_”, then it will show null.

2. Then add a conditional column, if it contains “_”, then it will show null.

3. We fill down the custom column. And filter the custom.1 is not null.

4. At last we removed the column1, and transposed the Table.

If it doesn’t meet your requirement, could you please show the exact expected result based on the table that you have shared?

Best regards,

Community Support Team _ zhenbw

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

BTW, pbix as attached.

[silverdale9999]

Hi - just dropping a quick message to say thanks.. I was on holidays last week so missed your reply.  I will give this a try and see where it takes me 😁

[v-zhenbw-msft]

Hi @silverdale9999 ,

How about the result after you follow the suggestions mentioned in my original post?

Could you please provide more details or expected result about it If it doesn't meet your requirement?

If you've fixed the issue on your own please kindly share your solution.

If the above posts help, please kindly mark it as a solution to help others find it more quickly. Thanks!

Best regards,

Community Support Team _ zhenbw

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[silverdale9999]

Hi - I really appreciate your continued support with this.

I really like what you've done and it's 100% new stuff for me which is great, but... It doesn't quite cover things.  Perhaps I should have explained it better, but *all* of the text is in one field at the start... I have uploaded an example file of how the data that is loaded might look:  https://1drv.ms/x/s!AjZk93r8SbJKg22k7Vh1vMJNWbN2?e=woKmml 

So each Change has it's own ref (which is unique) and a notes field with all the useful information in.

Thanks again!

[v-zhenbw-msft · Accepted solution]

Hi @silverdale9999 ,

We need to split the cell into rows first.

Select Note column and split it.

The remaining steps are the same as the reply above.

The result like this,

If you have any question, please kindly ask here and we will try to resolve it.

Best regards,

Community Support Team _ zhenbw

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

BTW, pbix as attached.