Hi

I need to do some thing like "find and insert before that " in a file which contains many records. This will be clear with the following example.

The original data record should be some thing like this

60119827  RTMS_LOCATION_CDR    INSTANT_POSITION_QUERY    1236574686123083rtmssrv7      20090309102806279           441           442
           783           WEB       1568          GMLC919443259137    FAILURE6         1236574665595654lessrv1

60119827  RTMS_LOCATION_CDR    INSTANT_POSITION_QUERY    1236574686123083rtmssrv7      20090309102806279                         442
           783           WEB       1568          GMLC919443259137    FAILURE6         1236574665595654lessrv1

How would you know where EXACTLY to insert the missing 'data'?
How do you know that the missing '441' should be surround by n spaces/tabs on the left and m spaces/tabs on the right?

Hello there,

As it was said, you really have to define a pattern/structure for your file including the exact field numbers where data is supposed to be modified or at least a group of specific values according to which the script has to search the file and to add what is missing.

Just for giving you a first idea, if the specific values are 441 and 442, then the following KornShell script will do the job

#!/bin/ksh

RESULT=""

while read LINE
do
    for ITERATOR in $LINE
    do
        if [[ $ITERATOR = "442"  ]]
        then
            RESULT="$RESULT 441 $ITERATOR"
        else
            RESULT="$RESULT $ITERATOR"
        fi
    done
    RESULT="$RESULT\n"
done < $1

print "$RESULT" > $1

Hi
where i can mention the file name to read the lines from...in ur script.

Obviously on the command line, for example if the script file name is myscript.ksh and your file name is myfile you write:

$ ./myscript.ksh myfile

Hi

Its working but there is realignment in the data spacing.

The existing data positions should not be disturbed and the number of characters 248 should be maintained in the result too

There will be blank spaces at the end of each line and in between data as well. the blank space indiactes there is no data for that field.

Actually there are certain field names by which the reports are generated, if that field is not available in the report there will be blank space to mention that.

I need to maintain the length 248 characters. Since this file will be sent for formatting to different format in another server. So data has to be intact.

there must be 11 blank spaces after the timestamp : 20090309102806279
after that i need to insert the 441 if 442 is present.

It's very difficult in script shell to manipulate line with spaces ... Try this code and launch it like : ./myscript infile outfile

Assume value 441/442 on positions 116 and 130.
If some lines contains a # character, change it in the script with another non-used character.

#!/bin/ksh

POS1=116 ; VAL1=441
POS2=130 ; VAL2=442

VAL_LEN=3

XPOS1=$((POS1 + $VAL_LEN - 1))
XPOS2=$((POS2 + $VAL_LEN - 1))

echo POS1=$POS1,XPOS1=$XPOS1
echo POS2=$POS2,XPOS2=$XPOS2

infile=$1
outfile=$2

NBL=$(wc -l $infile)
NBC=0

while [ $NBC -lt $NBL ]
do
        (( NBC = NBC + 1 ))

        xval1=$(head -$NBC $infile | tail -1 | cut -b$POS1-$XPOS1)
        xval2=$(head -$NBC $infile | tail -1 | cut -b$POS2-$XPOS2)

        if [ "$xval2" = "$VAL2" ]
        then
                if [ "$xval1" = "$VAL1" ]
                then
                        head -$NBC $infile | tail -1 >> $outfile
                else
                        # Must replace with $VAL1

                        echo "Must replace on line "$NBC

                        k1=$(head -$NBC $infile | tail -1 | sed 's/ /#/g')
                        (( ZPOS1 = POS1 - 1 ))
                        k2=$(echo $k1 | cut -b1-$ZPOS1)
                        (( ZPOS1 = XPOS1 + 1 ))
                        k3=$(echo $k1 | cut -b$ZPOS1-)
                        
                        echo "${k2}${VAL1}${k3}" | sed 's/#/ /g' >> $outfile
                fi
        else
                head -$NBC $infile | tail -1 >> $outfile
        fi
done