Tweak to sed/bash script needed

bulgin · 11-29-2019, 05:55 PM

Hello, see attached sample-data2.txt and bashy.sh script.

bashy.sh worked with other but very similar data to the attached sample-data.txt to enumerate the gene data enclosed in parenthesis.

Original data as $1:
./bashy.sh "Gene Code (4A - 4F) D2 fragment, D74F"

The data I modified for use with this script to enumerate addresses as $1:

./script.sh "1824 3rd st (1A - 1F) Brooklyn NY, 12550-4456"

and it worked!

However, now I need to:

rather than use $1 to bring in only one entry, I would like to bring in a txt with entries, such as sample-data2.txt

That's where I'm stuck.

(rename bashy.sh.txt to bashy.sh)

Turbocapitalist · 11-30-2019, 01:18 AM

It's easier if you use [code] [/code] tags around the script and post it inline. It will get more views that way.

That said, if I understand the question, you can wrap what you have as a function and then call the funtion on each line of the input file. Here is one way:

Code:

#!/bin/sh

gc() {
        PART1=$(echo "$1" | sed 's/\(.*\)\s(.*/\1/')
        . . .
}

while read g; do
        gc "$g"
done < $1

grail · 11-30-2019, 02:23 AM

Please show the desired output from sample-data2.txt

rnturn · 11-30-2019, 03:17 PM

Quote:

Originally Posted by bulgin

However, now I need to:

rather than use $1 to bring in only one entry, I would like to bring in a txt with entries, such as sample-data2.txt

That's where I'm stuck.

(rename bashy.sh.txt to bashy.sh)

If I'm interpreting your post correctly (which is questionable given the amount of cold medication I've been downing), what I would do is:

Take everything from "PART1=$(echo "$1" ..." to "done" and turn it into a bash function:

Code:

function parse_record() {

   PART1=$(echo "$1" ...

   ...

   done
}

Another assumption: You got that code block doing what you need it to do. (It sounded like that was working OK in your OP.)

Then, at the bottom of your script, use a loop like:

Code:

INPUTFILE=$1
cat $INPUTFILE | while read RECORD
do
   if [ -n "${RECORD} ]; then     # Process only non-empty records though we could
      parse_record "${RECORD}"    # get fooled by records with nothing but whitespace.
   fi
done

To further tweak the main script to handle multiple files on the command line, wrap the following around the above code block:

Code:

FILELIST=$*
for INPUTFILE in ${FILELIST}
do
   if [ -r ${INPUTFILE} ]; then    # File exists and we can read it.

      <see code block above minus the INPUTFILE=$1 bit>

   fi
done

Hoping this gets you close to what you're looking for.

Cheers...