Extracting multiple substrings from a text string

It is probably possible to create multiple child records, using Regex syntax in a pipeline, but that might be a little too fancy for both of us.

Assuming that there is some practical limit on the number of step strings like say 20 or even 50, it would be possible to use either a series of separate fields or one giant field with formula variables, and it would use the FIND function to locate the first occurrence coming out the string from the left. Then you could extract off the first one and then use the NotLeft function to hive off the string up to the end of the first stp string, then repeat the process like 20 times.

Each of the extracted strings would be in a separate form of variable, and then you would use the list function to list them all together, separated by a comma space 

I think you can jusy repeat this like 20 times.  This one does the first two.  You can just repeating the pattern.

var text StringOneStart = NotLeft([Stp String], Find([Stp String],"stp")-1);
var text StringOne = Trim(Left($StringOneStart, "|"));

var text StringTwoStart = NotLeft($StringOneStart, Find([Stp String],"stp")-1);
var text StringTwo = Trim(Left($StringOneStart, "|"));

List(", ", $StringOne, $StringTwo)

The issue with that approach is the number of "stp" strings is unknown.  If QB formulas supported looping it would be easy to solve.

If you want to venture into pipelines, you can achieve this with a simple regex and loop. 

If you have existing data you want to apply this to, start your pipeline with a 'search' step, then inside the loop just use an 'update record' step. And, once you update existing records and know that no other record has come during or after, then swap the search for on new event so you can update the records in near real-time. Before swapping the steps move the 'update record' step outside the loop otherwise it will be deleted when you delete the loop. 

Assuming you have a plan text field in your table to put the results to, you'll switch the entry of the filed to code using '</>' and input the below: 

*assumes your current field with this data is 'comments', we'll input our returned data into 'transformed comments'. Fields to update would be 'transformed comments' and then:

{% set match = a.comments | regex('(stp\w+)', true) %} 
{% if match is iterable %} 
{% for i in match -%}
{{ i }}
{%- if loop.last == false %}, {% endif -%}
{%- endfor %}
{% else %} 
No match
{% endif %}

** you may have to test for only a single 'stp...' value as we are iterating over the matched array.