Thread: bash search script improvement

hi,
i've switched ubuntu 10.10 few months ago , have started tinkering around bash scripting. ive been working on script searches filesystem specific file or directory, reason being find gui based search bit slow. have no prior programming/scripting experience, , ive written script want [ albeit imperfectly] in under 5 seconds.

problem search function imperfect, ie can identify directories, if last character of directory provided user. that.
also, don't know if ive commented script correctly [too much? little? ] , importantly, if code wasteful / plain bad.

along of these lines appreciated...

code:

#! /bin/bash  # script search specific file or directory in home folder , mounted filesystems, excluding /bin, /var , on... # disclaimer- know zilch, zero, nothing sed or awk. sed usage has been copied/modified google searches, , dont know how works  # [despite having  read few guides]... # whole sript built around grep searches, find compiling file lists, , wc -l line counts..  # function make index files, searched grep calls later locate files.... function makeindex {  echo "indexing files..." cd /media sudo find ./* -type f > /tmp/medialist.txt  cd /home sudo find ./* -type f > /tmp/homelist.txt  echo "indexing directories..." cd / sudo find . -type d > /tmp/dirlisttemp.txt  # add comma end of each line in dirlist, grep calls [later] can identify specific directory names # ie, grep searchterm returns "etc-searchterm-etc", grep "searchterm," returns searchterm, , not subdirs sed 's/\(.*\)/\1,/' < /tmp/dirlisttemp.txt > /tmp/dirlist.txt rm -f /tmp/dirlisttemp.txt  # read number of lines in index files determine total objects indexed, file sizes (stat) , report user set `wc -l /tmp/dirlist.txt` dirnum="$1" set `wc -l /tmp/medialist.txt` medianum="$1" set `wc -l /tmp/homelist.txt` homenum="$1" totalnum=$(($homenum+$medianum)) set `wc -l /tmp/dirlist.txt` dirnum="$1"  echo "$medianum files indexed in /media" echo "$homenum files indexed in /home" echo "$dirnum directories indexed" size=0 x in /tmp/homelist.txt /tmp/medialist.txt /tmp/dirlist.txt ;     set `stat -t $x`     size=$(($size+$2)) done sizef=$(($size/1000000)) echo "total index size $sizef megabytes" sleep 1  }  # check - missing command line arguments / argument -u regenerate index [ "$1" = "-u" ] && { echo "regenerating index files..." ; makeindex ; exit ; }   [ "$1" = "" ] && { echo "missing argument...." ; echo 'usage: search.sh "file/dir name" -d[irectory search]' ; echo "or search.sh -u rebuild index" ; exit ; } # assign $1 variable searchname, because value of $1 modified later in script cannot used search searchname="$1"  # check index files     ls /tmp/homelist.txt &> /dev/null || { echo "logfile not found...generating file index, may take while..." ; makeindex ; }         ls /tmp/dirlist.txt &> /dev/null || { echo "logfile not found...generating file index, may take while..." ; makeindex ; }     ls /tmp/medialist.txt &> /dev/null || { echo "logfile not found...generating file index, may take while..." ; makeindex ; }    #search directory matches if $2 -d     if [ "$2" = "-d" ];         echo "searching directory" # search user inputted value followed comma, subfolders ignored # however, means last character of dirname has provided user, ie # dir named "test1" cannot found cmd line input of "test"          grep -i "$searchname""," /tmp/dirlist.txt > /home/$user/dirsearch.txt         set `wc -l /home/$user/dirsearch.txt`          dirnum="$1"          echo ""$dirnum" directories found."         if [ "$dirnum" -ge "1" ];             echo "displaying results. results have been saved /home/$user/dirsearch.txt"             echo ""             cat /home/$user/dirsearch.txt         fi         echo "press enter continue searching in file lists or ctrl-c exit"         read userwaste     fi           echo "searching ""$1"" in index..."     rm -f /home/$user/searchresults.txt   # search files | ignore folders [crude code, needs improvement dont know how] ...     grep -i "$searchname" /tmp/homelist.txt | grep -i -v "/""$searchname""/" > /home/$user/searchresults.txt             grep -i "$searchname" /tmp/medialist.txt | grep -i -v "/""$searchname""/" >> /home/$user/searchresults.txt                   set `wc -l /home/$user/searchresults.txt`     matchnum="$1"      [ "$matchnum" = "0" ] && { echo "" ; echo "no results found" ; exit ; }      echo ""     echo "displaying results..."     echo ""         cat /home/$user/searchresults.txt     echo ""     echo "a total of $matchnum matches found"     echo "results saved /home/$user/searchresults.txt"

ps: i've searched on forum couldnt find better place post this. if im wrong, let me know , ill happily repost in correct section...

i took rough @ script. looks you're trying keep file , directory cache, search can done fast.

don't yet problem have script. post example of failed search?

regards.

Forum The Ubuntu Forum Community Ubuntu Specialised Support Development & Programming Programming Talk [SOLVED] bash search script improvement

Ubuntu