How do I decrement all array indexes in a text file?

user_8981474 0
Viewed 101

Background

I have a text file that looks like the following:

$SomeText.element_[1]="MoreText[3]";\r"
$SomeText.element_[2]="MoreText[6]";\r"
$SomeText.element_[3]="MoreText[2]";\r"
$SomeText.element_[4]="MoreText[1]";\r"
$SomeText.element_[5]="MoreText[5]";\r"

This goes on for over a thousand lines. I want to do the following:

$SomeText.element_[0]="MoreText[3]";\r"
$SomeText.element_[1]="MoreText[6]";\r"
$SomeText.element_[2]="MoreText[2]";\r"
$SomeText.element_[3]="MoreText[1]";\r"
$SomeText.element_[4]="MoreText[5]";\r"

Each line of text in the file should have the left most index reduced by one, with the rest of the text unchanged.

Attempted Solutions

So far I have tried the following...but the issue for me is I do not know how to feed it back into the file properly:

Attempt 1

I tried a double cutting technique:

cat file.txt | cut -d '[' -f2 | cut -d ']' -f1 | xargs -I {} expr {} + 1

This properly outputs all of the indicies reduced by one to the command line.

Attempt 2

I tried using awk with a mix of sed, but this caused by machine to hang:

awk -F'[' '{printf("%d\n", $2-1)}' file.txt | xargs -I {} sed -i 's/\[\d+\]/{}/g' file.txt

Question

How to I properly decrement all of the array indexes in the file by one and properly write the decremented indexes into the right location of the text file?

cat xargs sed substitution awk
3 Answers

A Perl one-liner makes this easy, overwriting the input file:

perl -pi -e 's/(\d+)/$1-1/e' your-file-name-here

(assuming the first number on each line is the index)

user_15773470

With simple awk you could try following, written and tested with shown samples.

awk '
match($0,/\[[^]]*/){
  print substr($0,1,RSTART) count++ substr($0,RSTART+RLENGTH)
}
' Input_file

OR in case your Input_file's count in between [..] is in any order then simply reduce 1 from them as follows.

awk '
match($0,/\[[^]]*/){
  print substr($0,1,RSTART) substr($0,RSTART+1,RLENGTH)-1 substr($0,RSTART+RLENGTH)
}
' Input_file
user_58665800

With GNU sed and bash:

sed -E "s/([^[]*\[)([0-9]+)(].*)/printf '%s%d%s\n' '\1' \$((\2 - 1)) '\3'/e" file

Or, if it is possible that the lines contain ' character:

sed -E "
/\[[0-9]+]/{
    s/'/'\\\''/g
    s/([^[]*\[)([0-9]+)(].*)/printf '%s%d%s\n' '\1' \$((\2 - 1)) '\3'/e
}" file
user_138090010
Related