Capture string in between two given string

[Niroj@Work]

Hi,

I need to capture a string between two given string using RegEx command.

Let>str1=ABC
Let>str2=ZZZ
Str2 may vary like "|" or "ZZ" r anything as user want.

Now,

Let>str=sshshcg ABC hghdg hgs 123 ZZZ hgshd gghgh hg ZZZ gchh

In the above string str, "ABC" is unique but "ZZZ" can occurs more than one. Here I want to extract data between "ABC" and 1st occurrence of "ZZZ" i.e( hghdg hgs 123) here.

Any idea..?

[Niroj@Work]

RegEx>%str1%(.+?)%str2%,%str%,0,,,1,$1,Val

Mdl>Val

Why this is not working?

[jpuziano]

Hi Niroj@Work,

If you want a purely regex solution, this can be done with lookahead and lookbehind... there's some info in this post:

RegEx> example using Lookahead and Lookbehind

Try this:

Code: Select all

Let>str1=ABC
Let>str2=ZZZ
Let>str=sshshcg ABC hghdg hgs 123 ZZZ hgshd gghgh hg ZZZ gchh
Let>pattern=(?<=%str1% ).+?(?= %str2%)
RegEx>pattern,%str%,0,match_array,num,0
MDL>match_array_1

The two spaces in the pattern are there so that the final result in match_array_1 does not have leading and trailing spaces.

If you want those leading and trailing spaces in the result string, just remove those two spaces from the pattern.

Does this work for you?

Take care

[Niroj@Work]

Thanks buddy! This is quite useful

[Niroj@Work]

I have different scenario where I need to replace the new line characters between two specific group.

For Example:

Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%

I need to replace all the new line characters between the curly braces {...}with null.

So required is:
Let>str=abcd %CRLF% hhjh jhj {select hgh hh=!3j h jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh nb jkj}%CRLF%

Could you please help here..

[Niroj@Work]

The below way I tried and replace the same at the end. But not able to replace at the same position. Whenever I am replacing {...} with _RPL_ME_ and later in the loop trying to replace the 1st occurrence I am getting absurd error.
======================
Program to Replace the same at the End:
================================
Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%


Let>pattern=\{.+?\}
RegEx>pattern,%str%,0,match_array,num,1,,str

Let>i=1

Repeat>i
RegEx>%CRLF%,match_array_%i%,0,,,1,,q
Let>str=%str% %q%
Add>i,1
Until>i>num

mdl>str


======================
Getting error: Program to Replace the {..%CRLF%...}at the same position:
================================
Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%


Let>pattern=\{.+?\}
RegEx>pattern,%str%,0,match_array,num,1,_REPLACE_ME_,str

Let>i=1

Repeat>i
RegEx>%CRLF%,match_array_%i%,0,,,1,,q
RegEx>(_REPLACE_ME_).+?,str,0,,,1,%q%,str
Add>i,1
Until>i>num

mdl>str

[jpuziano]

Interesting... I tried the following:

Code: Select all

Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%

Let>pattern=(?<={.+?)\r\n(?=.+?})

RegEx>pattern,str,0,match_array,num,1,~,new_string

MDL>new_string

However it fails with the error: Regular Expression pattern not compiled.

I'm not sure why though, maybe Marcus can shed some light.

Even if the above error could be eliminated though, I am not sure if this would catch and change multiple occurances of %CRLF% within { and } chars... it might only catch the first or last occurance within any { and } pair.

You can do this brute force with procedural code of course... but I am thinking you're looking for a regex solution here.

Marcus or anyone else, please jump in.

[Bob Hansen]

No access to Macro Scheduler right now, but does this work?

Let>pattern={(.*)%CRLF%(.*)%CRLF%(.*[^}])}

Replace with: $1$2$3

[Niroj@Work]

It will not work as %CRLF% may come any number times ..

[jpuziano]

Hi Bob,

Thanks for that, it got me on the right track.

Niroj@Work - try the following:

Code: Select all

Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%

Let>pattern=({.*?)\r\n(.*?)\r\n(.*?})

RegEx>pattern,str,0,match_array,num,1,$1%~$2~$3,new_string

MDL>new_string

I chose ~ as my replacement char instead of null so that the MDL> line can show the result to prove that it is working.

To replace with a null char instead, change the RegEx line to this:

RegEx>pattern,str,0,match_array,num,1,$1%NULLCHAR%$2%NULLCHAR%$3,new_string

Note though that then, MDL will not be able to show you the whole result string as it must interpret the first null char it hits as the end of the string and it stops there as in a "null-terminated string".

Does this work for you?

[jpuziano]

It will not work as %CRLF% may come any number times ..

I just saw this. If you're saying that %CRLF% may appear any number of times inside a set of braces { } then we'll need a different approach.

This one will be challenging to do with only RegEx... any other ideas anyone?

[Niroj@Work]

There is some problem with RegEx command itself.

Few days back I raised the issue with RegEx when we are matching a string containing curly braces.
http://www.mjtnet.com/forum/viewtopic.p ... highlight=

As curly braces are used for complex evaluation in macroscheduler it is creating problem.


Please run the below code:
=======================
Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%


Let>pattern=\{.+?\}
RegEx>pattern,%str%,0,match_array,num,1,_REPLACE_ME_,str

Let>i=1

Repeat>i
RegEx>%CRLF%,match_array_%i%,0,,,1,,q
RegEx>(_REPLACE_ME_).+?,str,0,,,1,%q%,str
Add>i,1
Until>i>num

mdl>str
==============

The error I am getting is: "Unknown Identifier SELECT" and "Unknown Identifier NB".. These are silly error is coming at "{" position.

[jpuziano]

Hi Niroj@Work,

There is something coming in Macro Scheduler version 12 that might address the problem that { and } chars may be causing due to their use in Complex Expressions. I'll leave Marcus to explain more about that...

In the meantime, I put together a non-RegEx solution that can handle any number of newlines within any { braces } section:

Code: Select all

Let>str=abcd %CRLF% hhjh jhj {select hgh hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jk %CRLF% j}%CRLF%
Length>str,str_len
Let>processed_str=
Let>char_located_inside_braces=NO
Let>char_index=0

Repeat>char_index
  Let>char_index=char_index+1
  MidStr>str,char_index,1,char
  If>char_located_inside_braces=NO
    If>char<>{
      ConCat>processed_str,char
    Else
      Let>char_located_inside_braces=YES
      //initialize this within-braces sub string
      Let>braces_section=
      ConCat>braces_section,char
    EndIf
  Else
    //code for when char_located_inside_braces=YES
    ConCat>braces_section,char
    If>char=}
      Let>char_located_inside_braces=NO
      //clean this within-braces sub string
      StringReplace>braces_section,%CRLF%,~,braces_section
      ConCat>processed_str,braces_section
    EndIf
  EndIf
Until>char_index=str_len

MDL>processed_str

Again, I chose ~ as my replacement char instead of null so that the MDL> line can show the result to prove that it is working.

To replace with a null char instead, change the StringReplace line to this:

StringReplace>braces_section,%CRLF%,%NULLCHAR%,braces_section

So... does this do the job for you?

[Niroj@Work]

It will not work as %CRLF% may come any number times ..

I just saw this. If you're saying that %CRLF% may appear any number of times inside a set of braces { } then we'll need a different approach.

This one will be challenging to do with only RegEx... any other ideas anyone?

As I mentioned this I am trying in the below way:
=============================
1) First I am trying to replace "{.....}" with "_REPLACE_ME_" anywhere in the line and catching the strings within "{...}" in match_array.

Let>pattern=\{.+?\}
RegEx>pattern,%str%,0,match_array,num,1,_REPLACE_ME_,str

2) Then I am replacing the %CRLF% in the strings by looping through "match_array" with NULL.
RegEx>%CRLF%,match_array_%i%,0,,,1,,q

3) After that again I am trying to replace the 1st occurrence of "_REPLACE_ME_" in %str% by %q% (But don't know how to do?).
I tried as:
Let>i=1
Repeat>i
RegEx>%CRLF%,match_array_%i%,0,,,1,,q
//Please try to replace _REPLACE_ME_ 1st occurrence.
RegEx>(_REPLACE_ME_).+?,str,0,,,1,%q%,str
//
Add>i,1
Until>i>num

And by doing so I am getting absurd error messages like unknown identifier SELECT, NB ... as I already discussed (in "{" issue).

Could you please try in my way once where for each value of match_array I want to replace the 1st occurrence of "_REPLACE_ME_" of %str% with %q%?

[Marcus Tettmar]

Interesting... I tried the following:

Code: Select all

Let>str=abcd %CRLF% hhjh jhj {select hgh %CRLF% hh=!3j h %CRLF% jh ? }jdjk jdkj jkj %CRLF% {nb bb hjdh %CRLF%%CRLF% nb jkj}%CRLF%

Let>pattern=(?<r>pattern,str,0,match_array,num,1,~,new_string

MDL>new_string

However it fails with the error: Regular Expression pattern not compiled.

I'm not sure why though, maybe Marcus can shed some light.

Even if the above error could be eliminated though, I am not sure if this would catch and change multiple occurances of %CRLF% within { and } chars... it might only catch the first or last occurance within any { and } pair.

You can do this brute force with procedural code of course... but I am thinking you're looking for a regex solution here.

Marcus or anyone else, please jump in.

Seems to only happen if the replace flag is set. If not replacing there is no error. I'm not sure why. I have put it on the to do list and will investigate.