I am using an array to store some data (keys e.g 47975081_1215781266128), it can be assumed that it is key to other data.

I want extract data from a file based on a couple of keys (range) and store the resulting data in a variable using the following command:
sed -n '/47975081_1215781266128/,/42628155_1215781428374/p' test1.txt

In my script I build a sed command and write it to a temp file using the code below:
Range1="${keys[$i]}"
Range2="${keys[$i+1]}"
a="sed -n '/"
b="/,/"
c="/p'"
d=" test1.txt"
echo "$a$Range1$b$Range2$c$d" > temp.txt

This works ok, it writes the following string to the temp file:
sed -n '/08465696_1215781522540/,/70225547_1215781581748/p' test1.txt

Then I grep this file and assign the contents to a variable
sedcmd=$(grep sed temp.txt)

However, the string that is extracted has other characters attached, the grep extracts:
sed -n ''\''/08465696_1215781522540/,/70225547_1215781581748/p'\''' test1.txt

It has added: ''\' before the '/08465.... and \''' after the ....48/p'
Can someone please tell me why? and how I can extract exactly what is in the temp.txt file and run it.

Basically, I want to build a command, run it and assign the resulting contents to an array.

I'm not seeing that.
Here are the contents of temp.txt:
sed -n '/08465696_1215781522540/,/70225547_1215781581748/p' test1.txt

Here is my script, test:
And I get the contents of the file in $sedcmd:
Are you sure that temp.txt has the right contents?

Thanksk, it was writing the quotes to the file. I have correct the problem. However, I am now faced with another for which there may be a simple explanation. When I run the command:

cat test1.txt | $($sedcmd) > temp2.txt

where sedcmd = sed -n '/08465696_1215781522540/,/70225547_1215781581748/p'

I don't get an output in the temp2.txt. When I run the script in debug it states 'No such file or directory'

The same message is displayed in debug even when I run:
sed -n '/08465696_1215781522540/,/70225547_1215781581748/p' test1.txt > temp2.txt


Any ideas why?

It will not be executed without backticks or simple brackets. Chechk the underlined backticks below.

The command appears to execute OK. See below I have attached the output from debug:

+ Range1=70225547_1215781581748
+ Range2=
+ a='sed -n /'
+ b=/,/
+ c=/p
+ echo 'sed -n /70225547_1215781581748/,//p'
++ grep sed temp.txt
+ sedcmd='sed -n /70225547_1215781581748/,//p'
+ cat test1.txt
++ sed -n /70225547_1215781581748/,//p
+ '<_05_1:MessageIdentifier>ERR:70225547_1215781581748</_05_1:MessageIdentifier>
' '</_05:ReceiveRequest><?xml' 'version="1.0"' 'encoding="UTF-8"?><error:Excepti
on' 'xmlns:error="hhtp/messages/exception/2006-06"' 'xmlns
:xsi="http://www.w3.org/2001/XMLSchema-instance"' 'SchemaVersion="1.0"><error:Na
me>ABC</error:Name><error:Code>001</error:Code><erroretail>An' error occured
whilst trying to process a routing request, see attached exception list for 'det
ails</erroretail><_05_1:MessageIdentifier>ERR:70225547_1215781581748</_05_1:MessageIdentifier>: No such file or directory

You have that input file text1.txt. Run the sed command on the shell onto it 1st without redirecting it to an outputfile. Just test this, then start redirecting it, if it works.
And please use the tags [ c o d e ] and [ / c o d e ] to have your code more eyefriendly for us, ty

The code below works:

Range1="${keys[$i]}"
Range2="${keys[$i+1]}"
a="sed -n /"
b="/,/"
c="/p"
echo "$a$Range1$b$Range2$c" > temp1.txt
sedcmd=$(grep sed temp1.txt)
errordata=$($sedcmd < test1.txt)
echo "$errordata" > temp2.txt

However, I am having to echo the contents of the variable to a file which I process further.

What I would like to do is send the data from the $sedcmd < test1.txt to temp2.txt by using: cat test1.txt | $($sedcmd) > temp2.txt. However, this gives me 'No such file or directory' errorr