Shell Programming and Scripting

Hi,

I would like to know if there is a command faster then sed, or a way to optimize my code...

Here is the description of my problem:
I have a file "file1" that is composed of 10 columns. The first one contain a list of ID, the second a list of number associated to the firs list of ID. The third one contains ID and the fourth the number associated, etc...

So I have:

Code:

ID  numb  ID  numb  ID   numb   ID   numb   ID   numb
ID  numb  ID  numb  ID   numb   ID   numb   ID   numb
...
...

The total number of ID is around 8000.

I want to find all the ID starting by a given string (eg. "460...") and extract the associated number. To do that I am simply using a grep command associated with cut and it is going pretty well.

Then I am doing an operation on the numbers extracted (MATLAB) and then I need to put the new values back in place in the file with 10 columns. To do that I am using:

Code:

x=file1
count=1
for k in $listID
do
  VALUE=`head -$count newval | tail -1`
  sed 's/'$k'  .\.....E.../'$k'  '$VALUE'/g' $x > extps
  \mv extps $x
  count=`expr $count + 1`
done

Where "newval" is the list of the new values computed, "listID" is the list of all the ID for which I need to change the associated number and "file1" is hte file with the 10 columns.

This code is working perfectly, but because I have a bunch of ID to change the repeted use of sed is taking A LOT of time (among 8000 ID I need to change about 500).

Does anybody know how could I speed up this code?

Thanks

This should be faster. Assuming the fields are separated by 2 spaces and you have the listID's separated with a space like something as below:

Code:

#!/bin/sh

listID="123 789"
VALUE="your_new_value"

awk -v l="$listID" -v var="$VALUE" '
BEGIN{FS=OFS="  ";split(l,arr," ");for(i in arr){a[arr[i]]=arr[i]}}
{for(i=1;i<10;i+=2)if($i in a){$(i+1)=var}}
{print}' file1

If you have the listID's line by line in a file (should be a better choice):

Code:

#!/bin/sh

VALUE="your_new_value"

awk -v var="$VALUE" 'BEGIN{FS=OFS="  "}
NR==FNR{a[$0]=$0; next}
{for(i=1;i<10;i+=2)if($i in a){$(i+1)=var}}
{print}' file_of_listIDs file1

If the output is correct you can redirect the output to a temporary file and replace file1 with the temporary file.

Use nawk, gawk or /usr/xpg4/bin/awk on Solaris.

Regards