Dividing tab blocks in bash script

Shell Programming and Scripting
1

Hi everyone,
I have a data.xml file which only contains thousands of data (tag) blocks. A part of the file looks exactly like this;

<data>
Line
Line
Line
</data>
<data>
Line
Line
Line
</data>

the rest of the file is simply a repetition of this part. Here each data block contains a number of multiple data lines. I need to separate each data lines so that it will look like;

<data>
Line
</data>
<data>
Line
</data>
<data>
Line
</data>
<data>
Line
</data>
<data>
Line
</data>

I need to write a bash script which reads data.xml and then executes the necessary operations. I guess the script file should have a loop with a conditional statement (if...then...else).

Any help will be highly appreciated.
thanks a lot.

2 
awk '/<data>/{f=1}/<\/data>/{f=0}f==1&&!/data/{printf "<data>\n%s\n</data>\n",$0;}' xml_filename

OR

awk '/<data>/{f=1;next;}/<\/data>/{f=0;next;}f==1{printf "<data>\n%s\n</data>\n",$0;}' xml_filename
1 Like
3

thanks a lot bipinajith, both works flawlessly :slight_smile:

4

Could you please explain step by step what you did there?

5

You can do it in sed -- read in second and third lines with N, N and if
open data, line, line
then make it
open data, line, close data, open data, line
then P out and discard three lines, N another line and branch back to if.

Else P out and discard one line and back to second N and if.

Do a $q before each N.

6 
awk '/<data>/{                          # If line contains pattern: <data>
  f=1;                                  # Set flag variable f = 1
  next;                                 # Skip current record
 } /<\/data>/ {                         # If line contains pattern: </data> - Since / is meta-character I escaped it \/
  f=0;                                  # Set flag variable f = 0
  next;                                 # Skip current record
 } f==1 {                               # If line flag variable is equal to 1 f == 1
  printf "<data>\n%s\n</data>\n",$0;    # Print <data> -- newline -- followed by current record : $0 -- newline -- followd by </data>
}' xml_filename                         # Read filename: xml_filename
7

I'd think you would need to print '</data>\n<data>\n%s\n",$0 to close out the first and put the second in a new tag. OK, you are discarding the old tags. That works.

8

Why do you use the next here?
Will this speed up the process some, since it does more than one thing in one run when finding the data tag?

9 
$ awk 'BEGIN {print "<data>"} /data/ {next} 1' ORS="\n</data>\n<data>\n" file | sed '$d'

or

$ awk '/data/ {next} 1' ORS="\n</data>\n<data>\n" file | sed '1 s/^/<data>\n/; $d'