[SOLVED] sed from constant regex

Issue

I tried to remove the unwanted symbols

%H1256
*+E1111
  *;E2311
   +-'E3211
      {E4511
       DE4513

so I tried by using this command

sed 's/+E[0-9]/E/g

but it won't remove the blank spaces, and the digits need to be preserved.

expected:

H1256
E1111
E2311
E3211
E4511
E4513

EDIT

Special thanks to https://stackoverflow.com/users/3832970/wiktor-stribiżew my days have been saved by him

sed -n 's/.*\([A-Z][0-9]*\).*/\1/p' file or grep -oE '[A-Z][0-9]+' file

Solution

You may use either sed:

sed -n 's/.*\([[:upper:]][[:digit:]]*\).*/\1/p' file

or grep:

grep -oE '[[:upper:]][[:digit:]]+' file

See the online demo

Basically, the patterns match an uppercase letter ([[:upper:]]) followed with digits ([[:digit:]]* matches 0 or more digits in the POSIX BRE sed solution and [[:digit:]]+ matches 1+ digits in an POSIX ERE grep solution).

While sed solution will extract a single value (last one) from each line, grep will extract all values it finds from all lines.

Answered By - Wiktor Stribiżew