Shell Programming and Scripting

hello,

first, yes i searched the forum , google and read many tutorials but still have a problem with my script.
I have great Problems, because i haven't worked with regular expressions before and never had anything to do with shellscripts. i am a complete Newby in this sort of theme.

I have many Files (end with .per) where RECORD should be renamed in Grid, thats easy with sed but now is the problem not everytime there is Record as text it should be renamed. It should only be renamed when after the RECORD(can be record with only small characters,too) if after the RECORD there is a blank(multiple blanks possible,too) then text(starts with a alphanumeric character but then there can be "_", Numbers in the text) .Following there must be a "[ one or 2 numbers ]".

EDIT:
I forgot to mention, that the the Files are not the same und die column and Row where it stands can differ from File to File, in the same file the searched pattern can exist multiple times. That means the script has to search each file from head to toe.

Examples:

Code:

RECORD s[11]
RECORD s_lt[6]
RECORD sn[9]
RECORD         f1035a[12]

This is the script i have written so far.

Code:

#!/bin/bash

for i in `ls *.per`; do

echo $i
awk '/ *[a-zA-Z]*\[0-9]|[0-9][0-9]\]/ {gsub(/RECORD/,"GRID");}' $i

done

Thank you for your help

Is your example file input?

Could you please show expected output from your example?

Does RECORD occur multiple times on one line? IF record is lower case do you want "grid" or "GRID"?

The example is how it is written in the file.
The Output should be:

Code:

GRID s[11]
GRID s_lt[6]
GRID sn[9]
GRID         f1035a[12]

The RECORD statement is only once per Line,
but as stated before can be in multiple lines in the same file.

I always want the uppercase "GRID"

Thank you for your help

I am not sure of the requirement , but based on the input and ouput you are providing i suggest :

Code:

 
sed 's/RECORD/GRID/g' input_file > output_file

That is not possible because in the Files are other Record statement
e.g

Code:

SCREEN RECORD f1051b(...)
SCREEN RECORD erfass    (...)

These RECORD differ ([] is not()) and are not allowed to get replaced, thats why the requirement is
RECORD name[number from 1-99]

Thank you for your Help

Based on your examples:

Code:

$ cat bla.txt
RECORD s[11]
RECORD s_lt[6]
RECORD sn[9]
RECORD         f1035a[12]
SCREEN RECORD f1051b(123)
SCREEN RECORD erfass    (456)
$ perl -pe 's/^RECORD(\s+[\w_]+\[\d{1,2}\])/GRID$1/' bla.txt
GRID s[11]
GRID s_lt[6]
GRID sn[9]
RECORD         f1035a[12]
SCREEN RECORD f1051b(123)
SCREEN RECORD erfass    (456)

An awk version:

Code:

 
 
awk '$0 ~ /\[[0-9]*\]/ {gsub("RECORD","GRID");print;next} {print}' input_file