Shell Programming and Scripting

Solaris? Try nawk instead of awk

Yep. Solaris bash

I used nawk and am not getting any errors now but the string is not being replaced.

Please see my updated mainfile.csv.

The field 26 has blankspaces and thus will not exactly match. Thats why i am using

Code:

[[ ]] and comparing with *

---------- Post updated at 01:37 AM ---------- Previous update was at 01:34 AM ----------

Can you kindly explain what you're doing here?

Code:

 'FNR==NR{C[$2]=$1;next}{$26=C[$26]}1'

If you don't mind the blank spaces being removed when field 26 is updated, this should work for you

Code:

nawk -F, '
FNR==NR{C[$2]=$1;next}
{sub(/ *$/,"",$26);$26=C[$26]}1' OFS=, reffile mainfile.csv

As for the explination:

FNR=Current file row number
NR=Global row number
when FNR==NR first file is being processed (ie reffile)

C[$2]=$1;next
Build a lookup array with refcode (2nd field) as index and replace value (1st field) as data

sub(/ *$/,"",$26)
replace any trailing spaces from field 26 (assumption is they are spaces and not tabs)

$26=C[$26]
lookup array with field 26 value and put data back into $26

The code is working now. Thanks a mil for the explanation. It makes a lot of sense now

---------- Post updated at 05:12 AM ---------- Previous update was at 01:52 AM ----------

I am trying to adapt the code so that the replacement is done only if the pattern in $26 matches the pattern C[$2] in the below code.

Where do i insert an IF loop like if ( $26 == C[$2] )

Code:

nawk -F, '
FNR==NR{C[$2]=$1;next}
{sub(/ *$/,"",$26);$26=C[$26]}1' OFS=, reffile mainfile.csv

Again you probably want to trim $26 before doing a compare so:

Code:

nawk -F, ' 
    FNR==NR{C[$2]=$1;next} 
    {sub(/ *$/,"",$26);if($26==C[$2])$26=C[$26]}1' OFS=, reffile mainfile.csv

No joy there.


When i insert the if condition there, the code doesnt work. Nothing is replaced meaning nothing is matched. C[$2] will have a different meaning since NR is now 2 right?