Filter using Regular expression under SMAT-View fr field location within segment

[Varun Sinha]

Hi,

I have successfully searched if hl7 messages contain email addresses (as *.com) in SMAT files via SMAT-View- Add Specified Messages To View- Add A Regular Expression Condition with this very simple regular expression (?s).*\.com.

As I see the documentation for Cloverleaf 6.1, it shows only content search.
Can we search based on HL7 field location within specific segment using SMAT-View?

e.g. Using regular expression under SMAT-View to search for messages having PID-14 not blank?

Can we do something like that?

 

https://usspvlclovertch2.infor.com/forums/topic/smat-regular-expression-compound-search/

 

Thanks & Regards,

Varun

Topic

[Charlie Bursell]

It could be done but would make for a very complex regexp.  Maybe you could dump to a file and examine messages there?

[Varun Sinha]

Thanks Charlie for the response!

Since, it’s production environment data, I don’t have luxury to dump it at a location. I can just search using SMAT-View.

Using regular expression under SMAT-View: I want to search for messages having PID-14 not blank?

The following regex under SMAT-View doesn’t work

\rPID(?:[^\|]*\|){5}(\|)

 

Regards,

Varun

[Charlie Bursell]

I told you it would be hairy.  The problem with using repetition meta characters like:

PID(\|.*?\|){14}

Is the counts are way off since you consume the | with each repetition so you have to make a long and sometimes unwieldy regular expression

The following regexp is true if field  5 is empty.  I’ll let you add additional up to 14.  Note the regexp will also not be true if less than x number of fields.

{\rPID\|.*?\|.*?\|.*?\|.*?\|\s+\|}

Good luck with it

[Charlie Bursell]

Got to thinking about this and though field 14 being empty would not necessarily be filled with a space.  So maybe the regex should be more like:

{\rPID\|.*?\|.*?\|.*?\|.*?\|(\s+|\|)}

I sure wish I had a SMAT file to play around with.

[Varun Sinha]

Thanks Charlie for your response!

 

But, I am getting following error while executing the below regex via SMAT-View- Add Specified Messages To View- Add A Regular Expression Condition

{\rPID\|.*?\|.*?\|.*?\|.*?\|\s+\|}

{\rPID\|.*?\|.*?\|.*?\|.*?\|\s+\|}

 

Error: The regular expression you typed is not valid.

 

Thanks & Regards,

Varun

[Charlie Bursell]

Let me do some more testing.  I don’t get an error but also do not get the desired results every time

[Charlie Bursell]

This thing has been bugging the heck out of me.  Finally decided I was getting spoofed by combining greedy and non-greedy expressions – a no no.

So I figured it out.  I ran some records through the engine and saved in SMAT so I could test.  This works!

 

\rPID\|([^|]*?\|){13}(\s+|\|)

This says look at the 14th field.  If empty or contains spaces it will match.  If less than 14 fields it will not match.

 

Try it.  Let me know.

[Varun Sinha]

Thanks a lot Charlie!

The above regex worked beautifully with desired result all the time. Thanks for your effort by looking into this issue.

Please let me know any best regex tutorial to crack these kind of scenarios within cloverleaf. How can we judge these regex against any cloverleaf versions for SMAT files and for SMATDB too?

 

Thanks & Regards,

Varun

 

[Charlie Bursell]

There are several good regex tutorials on Youtube and the web.  Some are using other programming languages but the information, except in a few cases, is the same.

I usually test my regexps by using a simple Tcl script with a lot of echo statements.  In the case of this one I simply created a file with a couple of HL7 messages, some that would pass, some that would not.  Then a simple Tcl script to test the regex using these files.  After passing in Tcl I tried it with SMAT.

The regex machine used by SMAT is in Java, I think.  There may be some nuanced differences but only with the most advance cases.

[Varun Sinha]

Thanks Charlie!

I am not very good at tcl script. Can you please share the tcl script, so that even I can test some regex.

Regards,

Varun

[Charlie Bursell]

There is no real Tcl script.  I simply run an interactive Tcl from the command line and test there.  if you want more complex you could write a simple script.

For example:

set str “

Hello, world!

”

regexp — {<(.*?)>(.*?)<} $str full part1 part2

echo $full
> 

Hello, world!<

echo $part1
> h1

echo $part2
> Hello, world!

If something specific or large I may read in str from a file.

As I said, there are may examples on the web and YouTube.  If another language the regexp is usually the same.  There are many Tcl examples as well  like:

https://wiki.tcl-lang.org/page/Regular+Expression+Examples

 

 

[Varun Sinha]

Thanks a lot Charlie for your help. I’ll try these out.

Once again, appreciate your help and guidance.

Regards,

Varun

 

[Author]

Replies