UNIX for Dummies Questions & Answers

Hi,

I am trying to parse a file of fixed length transactions into a new file. Using awk I need to add the line number as a 5 digit long numeric value to each of the transactional rows.

i.e.
awk '{print <ROW NUM PADDED HERE>substr($0,1,8).........}' filename > newfile

I know that NR-1 gives me the row number (as the file has a header record), however if I add NR-1 to the awk command this will give me records that are varying length i.e. the transaction on row 10 will be 1 charachter longer than the records on rows 1-9, and the records on lines in excess of 100 will be longer still.

Therefore, as part of the awk statement I need to pad the row number with leading zeros until it is 5 numbers long (the row number will never exceed 99999). I have read that using typeset -Z may be the way to achieve this, however using the syntax below only one leading zero is added to the row number, i.e. 01,02,03......010,011......0100,0101 etc. again this creates rows of varying lengths.

awk '{print typeset -Z5 NR-1substr($0,1,8).........}' filename > newfile

Can anyone help me with the correct syntax?

Thanks.

Code:

nawk '{printf("%.5d %s\n", NR-1, substr($0,1,8)).........}' filename > newfile

Hi,

Thanks for that, it seems to work fine although it puts an unwanted space between the row number and the next value printed out, for example if ($0,1,8) is ABCDEFGH then the output to the newfile is:

00001 ABCDEFGH

rather than:
00001ABCDEFGH

Any ideas?

Thanks

well......... it's kind of obvious if you looked at the solution - just remove the 'space' in 'printf'....

Code:

nawk '{printf("%.5d%s\n", NR-1, substr($0,1,8)).........}' filename > newfile