Get the middle sentence between words from email subject line

[ops_geo]

Hi I have created a simple flow where I extract the middle sentence from an email subject line. The subject line is usually in this format:

Attention - Department name, random words

I was able to extract the middle sentence using this formula in my initialize variable function:

first(split(last(split(outputs('Compose'),'- ')),',')) 

However, now I am finding that people are adding all sorts of "-" that is messing up the flow, for example, we have received emails with subject line that look like this:

1. Attention - Department Name, random words - more random words

2. Attention - Department Name - random words and more random words

I was wondering, is there a formular that will always extract words in between the first "-" and "," and ignore any subsequent "-" or ","

In some instances, people may put the Department Name in between two "-" "-" (see example 2 above).

How can I add that to my formula so that I can get the Department Name only?

If someone could please help me with that, I would much appreciate it. Thank you!

[srApps]

One way that worked in both of your examples was to do something like this where you can do it in two stages: first handle the "-" then handle the "," after.

Compose 2 is just this:

split(outputs('Compose'), '-')

This leaves you with this:

Compose 3 then checks to see how many times it split based on how many "-" there were in the string:

if(greater(length(outputs('Compose_2')), 2), trim(outputs('Compose_2')[1]), trim(last(outputs('Compose_2'))))

If it splits it into > 2 strings, get the 2nd with the index of [1], else get the last portion after the only "-".

Then, handle the "," in Compose 4:

first(split(outputs('Compose_3'), ','))

Should work as long as the department always comes after the first "-" and someone doesn't put another one before it.

[grantjenkins · Accepted solution]

This is how I'd probably extract out the department. See expression below:

trim(first(split(split(outputs('Compose'), '-')[1], ',')))

[srApps]

Concise, I like it. A simpler solution

[ops_geo]

@grantjenkins 

You area resident rockstar indeed!
quick question, what was the reason for removing "last" from the formula? Also, does [1] in bracket specifies the position of the item I want?

[grantjenkins]

If we work from the inside out.

trim(first(split(split(outputs('Compose'), '-')[1], ',')))

We use split to try and split on '-'. This will result in an array where we want the second item which is index 1 or [1]. Note that the first item at index 0 would be 'Attention ' in this instance.

We then use split to try and split on ',' and take the first item this time. We could have also just used [0] to get the first item.

Lastly, we use trim to ensure there are no preceding or trailing spaces around the Department name.

[ops_geo]

@grantjenkins 

makes sense, thank you so much for the explanation! I was wondering is the code only meant to work for "-" because I have noticed in emails that some - gets autoformatted to – (the longer version of the - which is so annoying because it breaks the flow, so I was wondering is there a way to consider both versions of the - and – ?

If you could please let me know, that would be great!