Issue in substitution

Shell Programming and Scripting
1

Hi ,

I have have file which has following structure

01aaaa88888000-9999
01ssss77777000-0991
01ssss7777700000991
02ssss7777700000991

The record 01 is corrupt as value from 12th field to 19th should be positive or start with - however it is 000-9999 it should be -0009999

i need to change the position of - to 12th field and add 0 in place of -
i will not touch 01 records which do not have - sign

Plz guide with approach

Thanks

2

You can use the following sed statement

sed -e "s/^\(...........\)\([^-]*\)-\(.*\)/\1-\2\3/g" input.txt

There are 11 dots in buffer 1 [surrounded by \( \)] which represent any 11 characters. The other two buffers contain the rest of the data without the - in them.

Instead of 11 dots, you can have

^\(.\{11\}\)

as well

3

Thanks a ton actually record length is 300 characters and error characters
0000-9999 is from 180th to 189th
plz guide in that scenario

4

I don't think sed supports the {##} suffix, but perl does. Here's an (untested) adaptation of the above:

perl -n -e 's/^(.{180})([^-]*)-(.*)$/$1-$2$3/' input.txt

js.

5

I have posted

sed -e "s/^\(...........\)\([^-]*\)-\(.*\)/\1-\2\3/g" input.txt

and

^\(.\{11\}\)

Do use the above two and create the sed statement you require.