Match multiple strings in multi lines and also replace multiple strings

Question

I want to replace all the strings with DISPLAY="TRUE" to DISPLAY="FALSE" in the first line and vice versa in the next line in a single match.

Example: FROM:

 <SYN DISPLAY="TRUE" SEARCH="TRUE" CLASSIFY="TRUE">Appels</SYN>
 <SYN DISPLAY="FALSE" SEARCH="FALSE" CLASSIFY="TRUE">103.103117.1031171012</SYN>

TO

 <SYN DISPLAY="FALSE" SEARCH="TRUE" CLASSIFY="TRUE">Appels</SYN>
 <SYN DISPLAY="TRUE" SEARCH="FALSE" CLASSIFY="TRUE">103.103117.1031171012</SYN>

Note that all other strings in the corresponding line of <SYN DISPLAY="TRUE" or <SYN DISPLAY="FALSE" could be different.

The requirement is to match and replace only in the occurrence of both lines (shown above) i.e when <SYN DISPLAY="TRUE" is in 1st line and <SYN DISPLAY="FALSE" is in the second line. Single lines with the following example pattern should not be replaced.

<DIMENSION_NODE>
            <DVAL TYPE="EXACT">
               <DVAL_ID ID="4294960976"/>
               <SYN DISPLAY="TRUE" SEARCH="TRUE" CLASSIFY="TRUE">2</SYN>
            </DVAL>
         </DIMENSION_NODE>
    ```

I tried using sed, however, I couldn't make it work.

sed -E 's/(<SYN DISPLAY=\")TRUE(\".+\s+<SYN DISPLAY=\")FALSE(\".+<\/SYN>)/\1FALSE\2TRUE\3/' test.xml

Requesting experts help to make it work :)

Why manually editing those lines will not be enough for you, Jerald? All you need to do is to modify two lines of a single file? What Note that all other strings in the corresponding line of ( means? — Quasímodo
– Quasímodo, Commented Apr 17, 2020 at 21:24
@Quasímodo There is 2000+ such '2 line' patterns in the file., so its difficult to manually edit . Sorry that the "note" part incomplete (because it was not quoted), I have corrected it now. — Devops Thrones
– Devops Thrones, Commented Apr 18, 2020 at 5:37
Is it possible that the first line has DISPLAY="FALSE" and the second DISPLAY="TRUE", and do they have to be changed? — Walter A
– Walter A, Commented Apr 18, 2020 at 10:36
@WalterA No that is not in the file when I checked. So only DISPLAY="TRUE" in the first line needs to be changed to DISPLAY="FALSE" and DISPLAY="FALSE" in the 2nd line needs to be changed to DISPLAY="TRUE", but there can be just single line with "DISPLAY=true", this is an XML file by the way. — Devops Thrones
– Devops Thrones, Commented Apr 18, 2020 at 10:56

Walter A · Accepted Answer · 2020-04-18 13:11:45Z

1

With -z the newlines will be handled as normal characters:

sed -zr 's/(SYN DISPLAY=)("TRUE")([^\n]*)\n([^\n]*)SYN DISPLAY=("FALSE")/\1\5\3\n\4\1\2/g' inputfile

In your example the remembered strings are:

\1=SYN DISPLAY=
\2="TRUE"
\3= SEARCH="TRUE" CLASSIFY="TRUE">Appels</SYN>
\4= <
\5="FALSE"

Both lines are used for the match, a single line will not be changed.

answered Apr 18, 2020 at 13:11

Walter A

20.2k2 gold badges29 silver badges46 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Devops Thrones Over a year ago

Thank you so much for the answer and explanation. This works as expected. One question, Is it possible to replace the group number (1 or 2, etc ) with some other string instead?

Walter A Over a year ago

@JeraldSabuM I do not understand your question. \1 should be used when you want to get the first remembered group, but you are not obliged to use that string. You are allowed to write

sed -zr 's/(SYN DISPLAY=)("TRUE")([^\n]*)\n([^\n]*)SYN DISPLAY=("FALSE")/syn comment=lowercase display=\5\3\n\4syn comment=otherChange display=\2/g' inputfile

Zorzi · Accepted Answer · 2020-04-17 21:13:20Z

1

From my point of view, the simplest way to achieve that, if you don't want to get into complicated regex, is to use a temporary value, like so:

sed \
    -e 's/DISPLAY="TRUE"/TMP/g' \
    -e 's/DISPLAY="FALSE"/DISPLAY="TRUE"/g' \
    -e 's/TMP/DISPLAY="FALSE"/g' YOUR_FILE

or, as a one-liner:

sed -e 's/DISPLAY="TRUE"/TMP/g' -e 's/DISPLAY="FALSE"/DISPLAY="TRUE"/g' -e 's/TMP/DISPLAY="FALSE"/g' YOUR_FILE

answered Apr 17, 2020 at 21:13

Zorzi

7925 silver badges9 bronze badges

1 Comment

Devops Thrones Over a year ago

Thank you for the answer, This works but the requirement is different, Sorry that I haven't made the question more clear. The requirement is to match and replace only in the occurrence of both lines i.e when <SYN DISPLAY="TRUE" is in 1st line and <SYN DISPLAY="FALSE" is in the second line because there could be single lines. For example, the below line alone should not be replaced: <SYN DISPLAY="TRUE" SEARCH="TRUE" CLASSIFY="TRUE">2</SYN>

Collectives™ on Stack Overflow

Match multiple strings in multi lines and also replace multiple strings

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related