UNIX for Dummies Questions & Answers

I would like to parse two strings from lines in a file only when both strings appear on the same line. For example, if I have the following line:

Code:

string1 string2 string3 string4 string5 string6 string7 string8 string9

I would like the output to be:

Code:

string2: string7

Can someone give me some advice? It may be possible that either string2 or string7 appear on a line, but not both and I don't want to parse those lines. I also don't want the text that appears between string 2 and string 7.

Thanks!

Is this a homework assignment? (If so, you must use the Homework & Coursework Questions forum and follow the special submission requirements for that forum.)

If not, I could interpret your request several ways.

Are you trying to print the 2nd and 7th "word" on a line for lines that contain 7 or more words?

Are you looking for lines that contain the literal string "string2" followed by the literal string "string7"?

Are you looking for lines that contain a word that is the literal string "string2" followed by a word that is the literal string "string7"?

Are you looking for lines that contain the literal string "string2" as the 2nd word on the line and the literal string "string7" as the 7th word on the line?

Is it one of the above but with words or strings matching two basic regular expression instead of two literal strings?

Please give us a more detailed description of what you're trying to do.

No this is not a homework assignment, this is work related. I am trying parse literal strings, not words, I just tied to simplify the example to make it easier to respond to. The strings are not necessarily the second or seventh word, but string2 always appears before string7. In fact, it is xml that I am trying to parse.

Thanks.

The following would seem to do what you're asking for as long as string2 and string7 don't contain any <slash> (/) characters or characters with special meaning in a BRE (<period> (.), <left-square-bracket> ([), <backslash> (\), <asterisk> (*), <circumflex> (^), and <dollar-sign> ($)):

Code:

sed -n 's/.*string2.*string7.*/string2: string7/p' file

Thanks Don. Actually string 2 looks like this:

Code:

(Interaction (username))

****where username can be any textual value

and string 7 is in the form:

Code:

TotalQuantity value="1"

****where 1 can be any numerical value

Andrew

By definition, "when username can be any textual value" and "where 1 can be any numerical value" do not describe literal strings. The following should do what you have described now if the version of sed you're using conforms to the standards:

Code:

sed -n 's/.*(Interaction (\([^)]*\))).*TotalQuantity value="\([0-9]*\)".*/\1: \2/p' file

If file contains:

Code:

<tag>(Interaction (user1))</tag><tag2>TotalQuantity value="121"</tag2>
<tag>Interaction (user2)</tag><tag2>TotalQuantity value="122"</tag2>
<tag>(Interaction (user3))</tag><tag2>TotalQuantity value="123x"</tag2>
<tag>(Interaction (user4))</tag><tag2>TotalQuantity value="124"</tag2>

the script above produces:

Code:

user1: 121
user4: 124

If you're using a GNU utilities version of sed and the above doesn't work, we can convert the BRE to an ERE and use awk instead of sed.