Hi Team I am getting db2 output as of below.

this    is    testing  1



this    is    testing   2





this    is      testing 4

The db2 provides output as, output for a query and a empty row and if there is no data for the query it provides one empty row

I want to remove the additional empty line after each output.

I know sed -i '/^$/d' file.txt can remove empty rows.
Is there a way to remove single empty row after each row which has data.

This is the desired output:

this    is    testing  1

this    is    testing   2



this    is      testing 4

This will only remove a newline if after some character (only the first newline after a line with text):

 sed -n '$!N;s/(.)n/1/;P;D' file

$ awk '{ print; if (length) getline }' file

this    is    testing  1

this    is    testing   2



this    is      testing 4

This prints the result and reads past the empty line after it if the result had any contents. For empty results, no additional line is skipped.

This would print multiple consecutive empty lines only if there was multiple consecutive empty results. The output would have exactly one line per result (empty or not):

$ cat file

this    is    testing  1



this    is    testing   2







this    is      testing 5



$ awk '{ print; if (length) getline }' file

this    is    testing  1

this    is    testing   2





this    is      testing 5

awk 'BEGIN {prev_empty = "TRUE"}; {if (/^$/) {if (prev_empty) {print}; prev_empty = "TRUE"} else {prev_empty = ""; print}}' file.txt

Explanation

• 
  BEGIN {prev_empty = "TRUE"}: start by initialising the variable prev_empty to TRUE.


• 
  if (/^$/): check if the line is empty (i.e. matches the regex /^$/).


• If so, then {if (prev_empty == "TRUE") {print}; prev_empty = "TRUE"}: if the previous line were empty, then print it, otherwise don't print (i.e. delete it). Then set the variable prev_empty to TRUE.


• If the current line is not empty, {prev_empty = ""; print}: set the variable prev_empty to nothing, and print the line.

The logic is that in almost all situations, print the current line. However, don't print the current line if the current line is empty (matches /^$/), and the previous line was non-empty (i.e. prev_empty = "").

If the file is small enough to fit in memory, you can slurp the file with perl and remove any newline that comes after a non-space character and another newline:

$ perl -0pe 's/(Sn)n/$1/g' file 

this    is    testing  1

this    is    testing   2



this    is      testing 4

With Awk, you could set the record separator to double-newline and then print each record with the default single newline output record separator.

With POSIX Awk, that will result in an extra newline after the last record - if you have GNU Awk (gawk) you can prevent that by setting an empty output record separator if the actual input record separator is empty:

$ gawk -vRS='nn' 'RT=="" {ORS=""} 1' file.txt

this    is    testing  1

this    is    testing   2



this    is      testing 4