3

It is pleasure to ask you about something I cannot solve myself. I am working with big text files with many similar lines. The format of the files is defined and I want to replace repetitively some characters appearing in some columns of the files. For example I would like to substitute in a file similar to this below, when the last number 3.0e-4 is, from the 5th to the 7th character by aaa, in this case yyy will be substitute by aaa:

ASDEFxxx          3.0E+12ERT          SAA      1.0e-4
ASEEFyyy          4.0E+12EDT          ABA      3.0e-4
ASAEFzzz          5.0E+12ERF          ADA      2.0e-4

In this example I would like to use replace regexp to replace in this case only the second line. The result should be:

ASDEFxxx          3.0E+12ERT          SAA      1.0e-4
ASEEFaaa          4.0E+12EDT          ABA      3.0e-4
ASAEFzzz          5.0E+12ERF          ADA      2.0e-4

There are many lines in the file to be substituted...

Does anybody write the command with replace regexp to do that?

Thanks and regards, Manuel

nanny
  • 5,756
  • 1
  • 20
  • 38
mnlmrn
  • 33
  • 3
  • You could write a macro. – Ista Jul 22 '15 at 22:02
  • 2
    (@Ista means keyboard macro, presumably.) – Drew Jul 22 '15 at 23:06
  • Yes, I meant record a keyboard macro. Thanks @Drew – Ista Jul 22 '15 at 23:49
  • Our of curiosity, how many lines is "big"? A keyboard macro won't really scale on the order of 100,000, and another method I'm thinking of won't scale beyond 1,000. (replace-regexp is the fastest method I know from a performance perspective) – PythonNut Jul 23 '15 at 03:55
  • Yes the file is about 500000 lines. I was exactly looking for this replace-regexp command. Thanks!!! – mnlmrn Jul 23 '15 at 14:11

1 Answers1

7

C-x h C-M-% \(.\{5\}\).\{3\}\(.*3.0e-4\) RET \1aaa\2 RET

Dissection:

  • C-x h select the whole file
  • C-M-% run query-replace-regexp (you may want M-x replace-regexp instead)
  • \(.\{5\}\) capture 5 characters into the group \1.
  • .\{3\} match three characters (will discard)
  • \(.*3.0e-4\) match the rest of the line, ensuring 3.0e-4 and capture in \2
PythonNut
  • 10,363
  • 2
  • 30
  • 76
  • Hi PythonNut, Many thanks for your answer. That is exactly what I was looking for.

    Thanks and regards Manuel

     – mnlmrn Jul 23 '15 at 14:09