Shell Programming and Scripting

This should be a really easy question.

My input file will have a few fields that are strings in the first line, which I will extract and save as variables. The rest of the fields on every line will be integers and floating point numbers. Can awk tell the difference somehow? That is, is there some way I can use if ( $i is not a string ) then {do stuff}? Or is that impossible? Thanks a lot!

This might help you understand ...
https://www.unix.com/shell-programmin...anumerics.html

Shows how to determine what kind of characters or numbers are present in a field.

Ahhhhh. Regular expressions are still kinda tricky for me, but such a powerful tool! I shoulda thought to turn there. Thanks for your quick help.

Be careful as -1.E2 is a number, too, but will not be matched by the regexes [0-9] nor [[:digit:]].
Adding +0 to force a field to become a number also can be misleading, as awk then will use any leading digits and ignore the residual non-digit characters. Still it may be promising to use sth like if ($1 == $1+0) do_stuff.

Fortunately, the project I'm working on will be simple enough that just checking for numbers should do the trick. However, * will also be used as a wildcard, and some numbers will be in the format xxx.xx.x.xx.xx or something similar. Should be a good exercise to teach me how to use regular expressions anyway!

You can just add '.' inside the list of characters to match '[...]', which BTW the list operator also escapes it from being a one-character-match-all metacharacter in all regex contexts. There are four or more flavors of regex, so in low/old and high/new contexts \< does not mean left side of word, and you need to use \b: Regex Tutorial - \b Word Boundaries Some grumpy old UNIX guys say POSIX team got railroaded by the PERL guys to change regex libraries to work their way. But one cannot live in the UNIX/LINUX/C/C++/JAVA world long without needing some regex experience.

Sometimes it gets kinky, like controls to make * shy not greedy (matches min pattern), similar to the difference between the ksh/bash ${name#pattern} (shy) versus ${name##pattern} (aggressive). One common way to make regex * less greedy is to say something like '[^;]*' not '.*', so it stops at the first ';'.