regex in sed to keep current line but remove next line

This is sample data:

ServerA
Value1 fh824rfz
Plan CustomA
ServerB
Value3 9fgjzxlo
Plan CustomD
ServerC
Value10 339fgh0l
Plan CustomE

This is the regex works in vscode:

(Value[0-9]{1,2} [0-9a-z]{8}\n)(.*)

Expected output:

ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

But I'm trying in sed with regexes like this but they don't work:

-E 's|(Value[0-9]{1,2} [0-9a-z]{8}\n)(.*)\n||g'
-re 's|(Value[0-9]{1,2} [0-9a-z]{8}\n)(.*)\n||g'
-zre 's|(Value[0-9]{1,2} [0-9a-z]{8}\n)(.*)\n||g'

How can I do this? I think the issue is with \n because when I remove that, the sample works (but still it's not the expected output).

Here's one way to do it (checked with GNU sed, syntax might vary for other implementations):

$ sed -E '/Value[0-9]{1,2} [0-9a-z]{8}$/{n; d}' ip.txt
ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

n command prints the pattern space (if auto-print is not disabled by -n option), replaces it with the next line and d will delete it.

By default sed is line-based and you cannot process several lines at once like you are trying to do in your two first atempts. To apply your strategy you must, for instance, concatenate the next line to the current one (N command) and then process this block of two line:

sed -E '/Value[0-9]{1,2} [0-9a-z]{8}$/{N;s/\n.*//}'

Another solution, if you are using GNU sed, if your file is not too large and does not contain NUL bytes (ASCII code 0), is to use the -z option to slurp the whole file as one single line with embedded newline characters (with the -z option sed considers that the input lines are terminated by a NUL byte instead of a newline):

sed -Ez 's/(Value[0-9]{1,2} [0-9a-z]{8}\n)[^\n]*\n//g'

This is almost what you tried with your 3rd attempt but as the (.*) group you used matches everything including newline characters you got only the first group printed... Replacing (.*) with [^\n]* in your 3rd attempt should work. Note also that as you want to suppress it there is no need for grouping.

With GNU sed you can also use the multi-line mode of the s command (m modifier) such that the period does not match a newline:

sed -Ez 's/(Value[0-9]{1,2} [0-9a-z]{8}\n).*\n//gm'

As others also proposed awk solutions here is a simple one:

awk 's{s=0;next} /Value[0-9]{1,2} [0-9a-z]{8}$/{s=1} 1'

That is, if variable s is different from 0 (or the empty string) reset it to 0 and skip the current line (next). Else, if the current line matches your regex, set variable s to 1. Finally, print the current line (1).

sed is usually not the best tool for anything that involves processing multiple input lines at a time.

Using any POSIX awk and only storing 1 line in memory at a time:

$ awk '!f; {f=(/^Value[0-9]{1,2} [0-9a-z]{8}$/)}' file
ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

Using GNU awk:

$ awk '{where=match($0,"Value[0-9]{1,2} [0-9a-z]{8}"); if (where) {print; getline} else {print}}' file
ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

With your shown samples please try following awk code. Using match function of awk which uses regex in it to get the matched values.

awk -v RS="" '
{
  while(match($0,/Value[0-9]+ [0-9a-z]{8}\n[^\n]*/)){
    val=substr($0,RSTART,RLENGTH)
    split(val,arr,ORS)
    prevLine=substr($0,1,RSTART-1)
    gsub(/^\n+|\n+$/,"",prevLine)
    print prevLine ORS arr[1]
    $0=substr($0,RSTART+RLENGTH)
  }
}
'  Input_file

Wow, all the professors have gathered here. Can I also suggest a way? 😉

awk --posix '/Value[0-9]{1,2} [0-9a-z]{8}/ { print prev ORS $0 } { prev = $0 }' data.txt

#OR

awk --posix '/Value[0-9]{1,2} [0-9a-z]{8}/ { print prev; print $0 } { prev = $0 }' data.txt

output :

ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

$ grep -v -f <(
     grep -P -A 1 "Value[0-9]{1,2} [0-9a-z]{8}" file |
     grep -P -v "Value[0-9]{1,2} [0-9a-z]{8}"
 ) file
ServerA
Value1 fh824rfz
ServerB
Value3 9fgjzxlo
ServerC
Value10 339fgh0l

save up $0 in awk, then use a dummy getline to "remove next line" by leveraging its return code as the substr()'s starting index, and using assignment to overwrite the "next line" with the current one :

jot 10 | 

mawk '(__ = $_)~/[3-7]/ && $!NF = substr(__, getline)'

3
5
7