Why do I need to double join a segment?

[Brandon Grudt]

As part of a project to strip as much sensitive data as we can, I’ve created an script that allows me to wipe multiple subfields.  My argument is {SUBS {PID.19.0}} -, but I can dump multiple period separated subfield locations into the single argument.

The pertinent portion of my script:

foreach swap $subs {

set swapL [split $swap “.”]

set swapsegL [split [lsearch -inline -all -regexp $segments [lindex $swapL 0]] $sep]

set swapfield [split [lindex $swapsegL [lindex $swapL 1]] $sub]

set newfield [join [lreplace $swapfield [lindex $swapL 2] [lindex $swapL 2]] $sub]

set newseg [join [lreplace $swapsegL [lindex $swapL 1] [lindex $swapL 1] $newfield] $sep]

set segments [lreplace $segments [lsearch $segments [lindex $swapL 0]*] [lsearch $segments [lindex $swapL 0]*] $newseg]

}

When I tested this the first time, my PID segment returned as a list, even though I had joined it by the separator.  I was able to fix it by modifying the above line to read:

set newseg [join [join [lreplace $swapsegL [lindex $swapL 1] [lindex $swapL 1] $newfield] $sep] “”]

It’s working now, but I was hoping to get clarification as to why I needed a second join to turn the PID segment into a string before placing it back into the segments var.  I’ve run into this before, and would like to understand why this happens.

Thanks in advance!

Topic

[David Barr]

Get rid of your “-all” argument to lsearch. This is making lsearch return a list of matching segments instead of the first one. You’re then treating this list as a string when you pass it to split.

[David Barr]

You also need to add a caret to your regex so that it only matches the segment name at the beginning of the segment and not within the data.

set swapsegL [split [lsearch -inline -regexp $segments ^[lindex $swapL 0]] $sep]

This code would be a lot simpler if you were using a message parsing library.

[Brandon Grudt]

The -all was in there for repeating segments, not that this would commonly be used for any repeating segments, but I always like to build for as broad a purpose as possible.  And David, do you have any information on using a message parsing library?

Thanks.

[David Barr]

If you want to handle repeating segments you have to use a loop, like this:

Code:

foreach swap $subs {
set swapL [split $swap “.”]
foreach segno [lsearch -all -regexp $segments ^[lindex $swapL 0]] {
set swapsegL [split [lindex $segments $segno] $sep]
set swapfield [split [lindex $swapsegL [lindex $swapL 1]] $sub]
set newfield [join [lreplace $swapfield [lindex $swapL 2] [lindex $swapL 2]] $sub]
set newseg [join [lreplace $swapsegL [lindex $swapL 1] [lindex $swapL 1] $newfield] $sep]
set segments [lreplace $segments $segno $segno $newseg]
}
}

If I was going to use a library, I’d write it something like

Code:

foreach swap $subs {
 lassign [split $swap “.”] seg field comp
 for { set i 1 } { $i <= [hl7::count hl7 $seg] } { incr i } {
   hl7::set_field hl7 ${seg}($i).$field.$comp ""
 }
}

You can get the library here: https://usspvlclovertch2.infor.com/viewtopic.php?t=4886

[Author]

Replies