Initial import of new posts

_posts/2007-05-06-useful-command-line-tips.markdown

@@ -0,0 +1,49 @@
+---
+author: einar
+comments: true
+date: 2007-05-06 08:40:21+00:00
+layout: post
+slug: useful-command-line-tips
+title: Useful command line tips
+wordpress_id: 244
+categories:
+- Linux
+---
+
+While working through large text files for work, I found several command line tools whose existence was unknown to me before. Their usefulness is great however, expecially if you deal with comma or tab delimited files (I do, and a great deal).
+
+<!-- more -->
+
+For example you can cut specific columns out of a file with the cut command:
+
+[code lang='c']
+cut -f [fieldname] textfile
+[/code]
+
+You can also specify ranges  (-f1,5) or a starting point till the end (-f2,. By default the field delimiter is TAB but you can change it with the -d option:
+
+[code lang='c']
+cut -d , -f1,2 textfile
+[/code]
+
+By default it outputs the result to the standard output, so you will need to redirect (>) to a file if needed. The --output-delimiter option permits you to change the delimiter of the output.
+
+The paste command does exactly the reverse:
+
+[code lang='c']
+paste file1 file2 > resultfile
+[/code]
+
+Basically it adds file2's lines to file1's, useful if you have to add a specific column from another file.
+
+Another nice utility is the "comm" command. I recently found it and it's a life saver if you have two files that contain similar elements and you want to find out which are common and which aren't. As a prerequisite the files should be **sorted** first, otherwise you won't get the right results. Then you can use comm to print out a series of information.
+
+[code lang='c']
+sort file1 > file1.sorted
+sort file2 > file2.sorted
+comm -13 file1.sorted file2.sorted
+[/code]
+
+In this particular example I tell comm to suppress output of the lines that are only in the first file (-1) and the common lines (-3) so that you effectively get only file2's items. The other option is -2, and that suppresses lines unique from file2. Again the output is on standard output so if you need to save it you'll have to redirect it somewhere.
+
+That's all for now. As you can see the command line can do a few interesting things (try that, cmd.exe) and that's merely the surface (google for other powerful commands such as awk or sed and you'll see what I mean). By doing things this way I manage my files far more efficiently than with a GUI.