Shell Scripting: how to pick value of an expression from each line of a file

I am new with shell scripting.

I am having a file containing some records of the form:

"text1:text2=value2,text3=value3,text4=value4,text5=value5"text1:text6:value6" "text1:text2=value2,text3=value3,text4=value4,text5开发者_运维知识库=value5"text1:text6:value6" "text1:text2=value2,text3=value3,text4=value4,text5=value5"text1:text6:value6" "text1:text2=value2,text3=value3,text4=value4,text5=value5"text1:text6:value6"

Now I want to write a shell script that picks up the value field for some text. Eg: I want value2 and value5 and I know that they will exist in front of text2= and text5=

Also there is no blank space in complete line. The file contains n lines and I want to have 2 values from each line(ie value2 and value5) and store then in a variables for further processing.

Can someone help.

Thanks

---

Using sed:

while read text2var text5var
do
    #something with text2var and text5var
done < sed 's/.*:text2=\([^,]*\),.*,text5=\([^"]*\)".*/\1 \2/') inputfile

Using GNU AWK (gawk):

while read text2var text5var
do
    #something with text2var and text5var
done < gawk -F ',|:|"' '{sub("[^=]*=","",$3); sub("[^=]*=","",$6); print $3, $6}' inputfile

To use other versions of AWK that don't have regular expressions for field separators, use a regex similar to the sed command or use a lot of splitting:

while read text2var text5var
do
    #something with text2var and text5var
done < awk -F ',' '{split($1,t2,"text2="); split($4,t5,"\""); split(t5[1],t5,"="); print t2[2], t5[2]}' inputfile

Using cut:

while read text2var text5var
do
    #something with text2var and text5var
done < cut -d , -f 1,4 --output-delimiter='=' inputfile | cut -d '"' -f2 | cut -d = -f1,3 | cut -d : -f 2 | cut -d = --output-delimiter=' ' -f1,2 

GNU cut may be required to be able to use the --output-delimiter option. It may be ugly but at least it's not being called four times on every line.

---

I'm sure that some more elegant solution is possible, but this bash script just loops through the input and filters out

• the value between the first = and the following , and

• the value between the fourth = and the following ":

  while read line
  do
      value2=`echo "$line" | cut -d = -f 2 | cut -d , -f 1`
      value5=`echo "$line" | cut -d = -f 5 | cut -d \" -f 1`
      echo $value2 - $value5   # do something with $value2 and $value5
  done
  

You call the script like this:

bash myscript.sh < mytextfile.txt

---

From the Command Line with text in q.text:

gawk -F\" '{print $2}' < q.txt | gawk -F: '{print $2 }' | gawk -F, '{print $1 "=" $4}'| gawk -F= '{print $2 "," $4}'

Tried on Cygwin bash and it will work. I am not a programmer but I use the cygwin shell and thought it would be fun to try doing this with gawk.