I think in this scenario I'd rebuild the geometry and then extract first and last coordinates rather than manipulating the text string

Edit: although it looks like this might not work for you if the decimal places are absolutely critical. In which case if you have already got a list you can use a list indexer with a list index of -1 to retrieve the last value

Hi @jez,

Another take is pure string manipulation:

Hope this helps.

Extracting the first and last coordinate pair from a string.fmw

You could use RegEx to get the precise characters.

Use the following expression in AttributeCreator or ExpressionEvaluator

@ReplaceRegEx(@Value(AttributeName),"\((\d+\.?\d* \d+\.?\d*),.* (\d+\.?\d* \d+\.?\d*)\)","\1,\2")

This will return Eg. "440720.53879545844 427232.77141405316,440720.07874009717 427234.3278727259 " being the first and last pair of coordinates in the WKT, with a comma delimiter.

Otherwise, the AttributeSplitter approach is fine as well, just use the ListIndexer in the way that @ebygomm suggests by retrieving Index = -1 to get the last value in the list.

@ReplaceRegEx(@Value(AttributeName),"\((\d+\.?\d* \d+\.?\d*),.* (\d+\.?\d* \d+\.?\d*)\)","\1,\2")

If only I knew how to write RegEx! This is so simple and works perfectly! Thank you

the Listindexer works perfectly too, I never thought to put a value of -1 in there to identify the last entry.

If you have time could you write out in english what the RegEx is doing - it all looks like smoke and mirrors to me.

OK, @jez it looks complex but really this is a simple one just using a few core RegEx expressions put together into a longer expression.

So really we start with the pattern we are looking for which is:

• A Left bracket "("
• + A number that may or may not have a decimal point and decimal component
• + A space character
• + Another number that may or may not have a decimal point and decimal component

....this is the first coordinate.

The last coordinate is similar, just with a right bracket, and at the end rather than at the start, so the pattern it obeys is:

• A number that may or may not have a decimal point and decimal component
• + A space character
• + Another number that may or may not have a decimal point and decimal component
• + A Right Bracket ")"

So put this together we have first coordinate as:

\\( = Left bracket. The "\\" is an escaping prefix we use because "(" is reserved in RegEx otherwise for special functions, so we need to tell RegEx that we literally mean the "(" character

\\d+ = A digit that is 1 or more characters long. This is what "+" as a suffix means

\\.? = A decimal character that may be 0 or 1 character long. Again "." is a reserved wildcard character in RegEx so we have to escape it with "\\" to tell it we literally mean the "." character.

\\d* = This is the decimal part following the decimal point "." that may or may not exist, and similarly the decimal number may be 0 or more characters long. This is what "*" as a suffix means.

,.* = This means the middle part of the string. This is a literal comma "," plus any combination of intermediate characters. "." is a reserved expression that means "any character" and ".*" means any character combination that is 0 characters or more long

So put it all together it just becomes (I'll just do the first half for the first coordinate. The last coordinate part is really just the same as below but with \\) at the end instead

\\( ----- \\d+ ----- \\.? ----- \\d* ----- {Space} ----- \\d+ ------ \\.? ----- \\d* matches

( ----- 440720 ----- . ----- 53879545844 ----- {Space} ----- 427232 ----- . ----- 77141405316

OK, now we introduce the special function of the "(" character into the mix. You put these around the parts of the pattern that you want to "Group" together as a substring. RegEx will "store" these Groups in memory as it parses through the string.

So we put (...) around the part of the pattern which corresponds to the numeric parts of the first coordinate Eg.

(-----(440720.53879545844 427232.77141405316) ----, = Group 1

(440720.07874009717 427234.3278727259) ----- ) = Group 2

With most Transformers in FME, you can "retrieve" those Groups of Substrings and call them as part of the expression of a new string, by Eg. using:

• \\1 to refer to Group 1
• \\2 to refer to Group 2

etc.

So you'll see in the RegEx expression the carefully placed extra brackets that create Group 1 and Group 2, and the Replacement string value we've chosen is just:

• \\1(Group 1, the bracketed first coordinate substring) + a comma character "," + \\2 (Group 2, the bracketed last coordinate substring)
• or "\\1,\\2" as the 3rd argument in the @ReplaceRegEx() function

You can try/learn about RegEx on quite a few online testers like https://regex101.com/ , which I find really useful for learning as it shows you how it applies each part of your RegEx expression to Test Strings.

I really appreacite your time with this. I will be copying the description you provided as I'm certain I will be able to utlise some of the functions in other tasks. I'll be visiting the website too.

You have been a great help. Thank you