Some Simple Shell Script Examples

Cut a block of lines

BatchFtp

Compute Average, Maximum and Minimum

Spelling Related Tools

Scripts for HTML

FAQ

FAQ2

---

Some Simple Shell Script Examples

☉ WARNING:

─ Exception handling are all omitted in the following examples.

─ Use with caution.

☉ Grep pattern with line number

─ Usage

gn [options] pattern files

─ Script

grep -n $*

☉ Check Mail: See the tail of the latest mail

tail /usr/spool/mail/uname

tail $1 /usr/spool/mail/uname

☉ Double space a text file

sed G filename

☉ Count the number of lines

wc -l $1 | cut -c1-7

☉ Generate a sequence of numbers

count=1 while [ $count -le $1 ] do echo $count count=`expr $count + 1` done

head -$1 longfile|cat -n|cut -c1-5

head -$1 longfile|awk "{print NR}"

head -$1 longfile|perl -p 'print $.'

☉ search a file recursively

find $HOME -name $1 -print

find . -name $1 -print

☉ convert a file into 2 coulumn

pr -t -l66 -w80 -2 $1

☉ Eliminate double space effect

─ when upload a file from PC

ex $1 <<% ___________ 1,\$s/^M\$// |1,$s/^M// | \$s/^^Z\$// |$/^Z// | w |w | % -----------

☉ change file names for many files

ls | sed 's/..*/mv & &/' > script vi script sh script

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  simple.htm, 

Cut a block of lines

☉ cut a block of lines

─ Usage

gl beginline endline file

─ Script

sed -n -e "$1,$2p" $3

QuitPoint=`expr $2 + 1` sed -n -e "$1,$2p ${QuitPoint}q" $3

☉ delete a block of lines

─ (dl startline endline filenmae)

cp $3 tmp.o echo "$1,$2d w" > script ex $3 < script

☉ replace a block of lines

─ (rl startline endline file1 file2)

mv $3 tmp.o a=`expr $1 - 1` b=`expr $2 + 1` head -$a tmp.o > $3 cat $4 >> $3 sed -n -e "$b,\$p" tmp.o >> $3 !!!!! What if $1 is 1 or $2 too long?

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  block.htm, 

BatchFtp

☉ batch ftp a file

ftp $1 <<% lien binary put $2 (OR get $2) quit %

☉ batch ftp many files

echo "ftp $1 <<% lien binary" > script for i do echo "put $i" done >> script echo "quit %" >> script sh script

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  batchftp.htm, 

Compute Average, Maximum and Minimum

☉ Compute average from a colume of a file

─ Usage

avg column file

─ Script

awk " {sum += \$$1;} END {print sum/NR} " $2

☉ get the maximum of any column

─ Usage

max colume filename

─ Script

awk " BEGIN { MAX = -999999} { if ( \$$1 > MAX) MAX = \$$1 } END {print MAX} " $2

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  avgmax.htm, 

Spelling Related Tools

☉ Dictionary look up

─ Usage

look starting_word_segment

─ Script

egrep "^$1" /usr/dict/all

☉ Grep with multiple patterns

─ Usage

mgrep pat1 pat2 pat3 ...

─ Script

case $# in 0) cat ;; 1) cat | grep $1 ;; 2) cat | grep $1 | grep $2;; 3) cat | grep $1 | grep $2 | grep $3;; esac

☉ Find an English word

─ Usage

e.g. looking for the word "convenient"

sphelp con v en t

─ Script

pat=$1; shift grep "^$pat" /usr/lib/dict/all | mgrep $*

pat=$1; shift look "$pat" | mgrep $*

☉ spelling corrector

spell $1 > 1.o comm -23 1.o excluedwords > 2.o sed 's/..*/1,$s\/&\/&\/g' 2.o > script vi script ex $1 < script

─ Pitfall

May hit unwanted words

1,$s/th/the/g w

Ex is atomic - Do all or do nothing

following script will fail

1,$s/systen/system/g 1,$s/systens/systems/g w

☉ Other way to do the 3rd line

(while read w do echo "1,\$s/$w/$w/g" done ) < 2.o > script

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  spell.htm, 

Scripts for HTML

☉ Generate a Color Table

echo "" for R in 00 33 66 99 cc ff do echo "<TABLE>" for G in 00 33 66 99 cc ff do echo "<TR>" for B in 00 33 66 99 cc ff do echo "<TH bgcolor=#$R$G$B> <font size=3 color=BLACK>$R $G $B<br>" <font size=4 color=#$R$G$B>$R $G $B</TH>" done echo "</TR>" done echo "</TABLE>" done

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  html.htm, 

FAQ

☉ 分割檔案 split

─ 例如將一個 2mb 的檔分成 2 個 1mb 的檔

split -b 1m filename

☉ 要如何改一堆檔名?????

─ 將*.foo改成*.bar

for i in *.foo do mv $i `basename $i .foo`.bar done

─ 將foo.*改成bar.*的方法

for i in foo.* do tailname=`echo $i | sed -e 's/^foo\.//'` mv $i bar.$tailname done

☉ 檔案截取命令

awk '/start-here/,/end-here/' data.rec

─ 要取出start-here到end-here之間的段落. 不包括 start-here以及end-here這兩行

用tail和head

─ 用awk

awk '{if ( $0=="start-here") \ n=1 \ else { if($0=="end-here") \ n=0 \ if (n==1) \ print}}' data.rec

☉ delete file when its size = 0 byte

find / -size 0 \(-name \*.dat -o -name sh\*.dat \) -exec rm -f {} \;

☉ Sort and Rename Files

#!/bin/sh for i in 000??.txt.all do sort -r $i > `echo $i | sed 's/\.all$/.sort/'` done

☉ 如何讓程式,在logout以後還能繼續執行

nohup command [ arguments ] &

☉ How to make a "script" to judge which the "shell" we are in?

echo $SHELL

☉ 將一個執行檔放到unix上時 要打什麼才能使用那個執行檔呢??

放上去後.. 打執行檔的名稱... 就會出現...command not found...

將一個執行檔放上Unix以後, 第一應該先檢查premission 可以先 chmod 755 <your file> 然後再用 ./ 執行,

其實有些 OS 預設就把 . 加入path裡, 所以不必加 ./

加個./其實是好習慣.. 當你到了別台設定不一樣的機器上 不會頓時不知所措

☉ Delete all lines that contain some string

我有一堆ASCII文字檔案，有沒有哪一個指令可以 搜尋到檔案中某一行字串，並且將該行刪除。 若是用VI，因為我的檔案多且檔案大，相當費時

grep -v "string" filename > new_filename

☉ How to use sed to add a character "." at the beginning of all lines in a file ?

sed -e 's/^/\./'

☉ 列印檔案(程式檔)時 如何讓printer 每一頁底下飲出頁碼呢??

Use "pr" to convert your ascii file first, then print it.

use "man pr" to know the details.

☉ 用perl 處理中文字, 要去掉那些特殊字元才不會有亂碼?

☉ Delete files with odd names

─ Problem

我做網站上傳檔案'後來改名稱'結果檔名變成前一個為空白就是 _INDEX.HTM-前面那條線是空白'無法刪除'

─ Solution

先移到跟那個檔案同一個目錄下 然後試試rm ./檔名吧 如果不可以的話....才試試rm -ir .(這樣比較危險一點) rm -r .的意思是砍掉目前目錄以及其下的子目錄及檔案 加上i的話,系統會在砍掉目錄或檔案前先詢問y or n 回答y才會砍,回答n就不砍,你看到要砍的那個檔案再按y 其他都按n(小心不要按錯了,砍了救不回來的)

☉ ftp 抓子目錄..

─ unix 有沒有甚麼ftp軟體 可以抓子目錄阿...

ncftp:

ncftp> get -R dir_name

wget. GNU software.

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  faq.htm, 

FAQ2

☉ Rename all the files in the currect directory to the lower case

for file in * do lcfile=`echo $file | tr "[A-Z]" "[a-z]"` mv $file $lcfile done

☉ A ex script

─ I have a file with data as such :

9600012301FDTF02FTDT03FFTF04TFDD05TDFF06TTTT*

seperate this data into the format :

96000123 0p1FDTF 0p2FTDT 0p3FFTF 0p4TFDD 0p5TDFF 0p6TTTT

sed 's/\([0-9][0-9][A-Z][A-Z][A-Z][A-Z]\)/\ \1/g'

It reads "insert newline before every occurrence of two numbers followed by four uppercase letters."

☉ Remove a blank line in a file

egrep -v '(^$)' test.txt

If spaces/tabs on a line count as a blank line, then you need a more complicated expression.

☉ Switch first two columns

─ sed/vi

:%s/^\([^ ]*\) \([^ ]*\)/\2 \1/

─ awk

awk '{print $2 $1}'

☉ An AWK problem

I am looking through a bunch of files in a directory. The files I am trying to isolate contain the "046*" in the 3rd record and the string "TOWER GROUP" in the 6th record. There are many files to go through and I have tried working with 'grep -l' but that only works in an ideal situation. The command 'awk' is useful but how do I display the contents of the entire file if the conditions are met.

BEGIN { FS = "," } $3 ~ /046\*/ && $6 == "TOWER GROUP" { while( (getline < FILENAME) > 0) print close(FILENAME) }

☉ Question : Comparing Number

I'm trying to compare two numbers. If the day of the month is less than 10 then echo "larger". Otherwise, echo "smaller".

#!/bin/sh num=date" +%d" echo "\n $num" if expr "$num" \> 10 then echo "larger" else echo "smaller" fi

☉ Filtering words

sed "s/this//;s/is//;s/a//;s/ *//g"

I'm trying to write a script to filter out (remove) certain words

─ eg.

I want to filter out: is,this,a Input: this is a dumb question Desired Output: dumb question

if you want to filter 'is' but not "isn't", that wouldn't work; you could still use sed, but I think I'd use perl so that I don't have to code 3 sed substitutions for every word:

perl -p -e 's/(^|\s+)is(\s+|$)//g;s/(^|\s+)this(\s+|$)//g;'

perl -p -e 's/(^|\s+)(is|this)(\s+|$)//g;'

☉ Finding files less than a day old

Use GNU find with -mmin option

find / -amin +120 ...

find . -newer thefile -print

find . -mmin +120 -type f -exec rm -r {} \;

國立政治大學資訊科學系 連耀南 lien@cherry.cs.nccu.edu.tw  faq2.htm,