[SOLVED] A little help with awk, just the first number- if 1 then 0 else skip it.

BW-userx · 07-04-2017, 11:08 AM

I got a bunch of image file with the numbering system as such
Image-113-1301.png
Image-113-1302.png
and others with
Image-013-1302.png
Image-013-1303.png

all I want to work with an change is the first set of digits with the leading number, or first number being
if 1 then 0 else skip it.

I've got awk to change the first instance of a 1 to 0, but that is not what I need.

Code:

$fix"Image-113-1301.png"  
$fixed=$(echo $fix | awk 'NR==1,/0/{sub(/1/,"0")} 1')                             
$echo $fixed                                                                      
Image-013-1301.png

that above one works but if the 1 is the second digit or more then it wil change that one instead, which I do not need. It needs to just do nothing to it.

Code:

$fix"Image-013-1301.png"                                                         
$fixed=$(echo $fix | awk 'NR==1,/0/{sub(/1/,"0")} 1')                             
$echo $fixed                                                                      
Image-003-1301.png

so I can rename all of them that have the leading number 1 to a zero before I can move on.

danielbmartin · 07-04-2017, 11:28 AM

I'm not sure that I understand the problem. Maybe this will do.

With this InFile ...

Code:

Image-113-1301.png
Image-113-1302.png
Image-013-1302.png
Image-013-1303.png

... this awk ...

Code:

awk '{sub(/Image-1/,"Image-0"); print}' $InFile >$OutFile

... produced this OutFile ...

Code:

Image-013-1301.png
Image-013-1302.png
Image-013-1302.png
Image-013-1303.png

Daniel B. Martin

Turbocapitalist · 07-04-2017, 11:35 AM

I'm not sure I understand either but about looking for a pattern at the start of a line, you need to anchor it with a caret ^

Code:

s/^0/1/;

Otherwise, without the anchor, the pattern will find any 0 anywhere in the string.

BW-userx · 07-04-2017, 11:58 AM

Quote:

Originally Posted by danielbmartin

I'm not sure that I understand the problem. Maybe this will do.

With this InFile ...

Code:

Image-113-1301.png
Image-113-1302.png
Image-013-1302.png
Image-013-1303.png

... this awk ...

Code:

awk '{sub(/Image-1/,"Image-0"); print}' $InFile >$OutFile

... produced this OutFile ...

Code:

Image-013-1301.png
Image-013-1302.png
Image-013-1302.png
Image-013-1303.png

Daniel B. Martin

thanks that worked!

Code:

#!/bin/bash

working_dir=/media/data/FIX-Numbering

while read FILENAME
do

f=$FILENAME
path=${f%/*}
xfile=${f##*/}
title=${xfile%.*}
ext=${xfile##*.}
fix="$(echo "$title" | awk '{sub(/Image-1/,"Image-0"); print}' )"
#test first
#echo "$fix"
mv -v "$FILENAME" "$path/$fix"."$ext"
done< <(find "$working_dir" -type f)

BW-userx · 07-04-2017, 12:06 PM

Quote:

Originally Posted by Turbocapitalist

I'm not sure I understand either but about looking for a pattern at the start of a line, you need to anchor it with a caret ^

Code:

s/^0/1/;

Otherwise, without the anchor, the pattern will find any 0 anywhere in the string.

thanks I was pattern thinking only the first number (digit) in the line
if = 1 then change it to zero (0)

a-11134-1123023

becomes

a-01134-1123023

that ^ looks good, I am just not that accustom with syntax of sed and awk (yet).

got it though thanks to the other poster. ! cheers.

scasey · 07-04-2017, 02:31 PM

Quote:

Originally Posted by BW-userx

I am just not that accustom with syntax of sed and awk (yet).

It's less about sed and awk (or perl or grep) than it is about "regular expressions" (regex) -- the power of those utilities lies in understanding how regular expressions work and the pattern matching therein. Web searches and study will help, but the learning curve is medium steep. Still, once you have your head wrapped around them, all those utilities become "easy" to use.

One thing I did was use grep to spin through web and email logs to understand how patterns get matched. Another great tool for me was in the SciTE editor, which has a "Regular Expression" option on it's Find and Replace functions. Very useful to have it highlight a regex find to see I did (or didn't) have the regex as I wanted it.

Learn regex first, then go to sed or awk or grep or perl to learn the syntax for applying them.

BW-userx · 07-04-2017, 02:51 PM

Quote:

Originally Posted by scasey

It's less about sed and awk (or perl or grep) than it is about "regular expressions" (regex) -- the power of those utilities lies in understanding how regular expressions work and the pattern matching therein. Web searches and study will help, but the learning curve is medium steep. Still, once you have your head wrapped around them, all those utilities become "easy" to use.

One thing I did was use grep to spin through web and email logs to understand how patterns get matched. Another great tool for me was in the SciTE editor, which has a "Regular Expression" option on it's Find and Replace functions. Very useful to have it highlight a regex find to see I did (or didn't) have the regex as I wanted it.

Learn regex first, then go to sed or awk or grep or perl to learn the syntax for applying them.

I do understand that is is /search pattern/replace pattern

it is not the first occurrence that matches then change, I was looking for the first char in a pattern that is a one to be changed among a bunch of data that has the same like pattern but different numbers - that is what got me. for example

image-110-1234
image-110-1235
image-001-1234

I was thinking- just looking at the entire string then moving down it left to right one char at a time until it sees the first number in the string then checking just that one digit -- is it a '1' one or other than a one?

if '1' then change it to '0' if not '1' then skip it goto next file line of thought I was stuck on.

it was more of a seek and destroy then a pattern match mind set I was using.

then I seen that Image-1/Image-0 then my brain just knew that was the answer I was needing.

scasey · 07-05-2017, 12:05 PM

I suppose there is a regex way to identify the first '1' in a string and then replace it. I've not had much luck with that. When I try to identify the first instance of a character, the match usually includes all instances. I've never figured out how to do that and just usually resort to something like the solution here. So, to be clear, while I'm saying that the power of sed/awk/grep lies in understanding regex, I'm not claiming to be an expert with it; I learn more each day

szboardstretcher · 07-05-2017, 12:15 PM

Python makes it easy to replace the first occurrence of a thing:

Code:

>>> import re
>>> filename = "Image-113-1301.png"
>>> re.sub("1","0",filename,1)
'Image-013-1301.png'

astrogeek · 07-05-2017, 02:08 PM

Quote:

Originally Posted by szboardstretcher

Python makes it easy to replace the first occurrence of a thing:

Code:

>>> import re
>>> filename = "Image-113-1301.png"
>>> re.sub("1","0",filename,1)
'Image-013-1301.png'

But that would change the first '1' in every filename, whereas the stated constraint is...

Quote:

Originally Posted by BW-userx

I got a bunch of image file with the numbering system as such
Image-113-1301.png
Image-113-1302.png
and others with
Image-013-1302.png
Image-013-1303.png

all I want to work with an change is the first set of digits with the leading number, or first number being
if 1 then 0 else skip it.

For that you need to provide some context for the '1' you want to change which is provided by the leading 'Image-' part, among other ways. Context is the idea behind regex anchors, ^=first of string, $=end of string, and user defined leading/trailing characters.