Home / Questions / 553840
Accepted
ccfz
Asked: 2022-08-27 01:05:59 +0800 CST 2022-08-27 01:05:59 +0800 CST

Replace XML tag text with shell

  • 772

I'm trying to replace a text string in an XML file with another using the sed tool command , but I can't get the result I expect.

Command:

sed -i -e -r 's|test-plugin.+\s\s*?<version>(.+)<\/version>|test-plugin.+\s\s*?<version>1.0.0<\/version>|g' plugins.xml

XML:

      <plugins>
        <plugin>
          <name>test-plugin</name>
          <version>3.0.0-SNAPSHOT</version>
        </plugin>
        <plugin>
          <name>test-plugin-2</name>
          <version>3.0.0-SNAPSHOT</version>
        </plugin>
        <plugin>
          <name>test-plugin-3</name>
          <version>2.0.0-SNAPSHOT</version>
        </plugin>
      </plugins>

The result I expect is that it replaces the version of the first plugin since it is the one that matches the name given in the command itself. Obviously, replace it respecting the line breaks and spaces of the XML structure.

Thank you very much in advance.

xml

2 Answers

  1. Best Answer

    Right now I can't think of a shorter version with awk, but it would definitely sedbe an unnecessary pain.

    Try this:

     $ awk '/<name>test-plugin<\/name>/{seen=1;print ; next}; seen {gsub(/3.0.0/, "1.0.0"); seen=0};1' archivo.xml

    What I do here is that when I find the name of the plugin, I assign a 1 to the "seen" variable (that is, I set it to True), then I print the line and ask awkit to see the next line, and then, if variable seenis True (has the value 1), substitute the version number and then set the variable to false seen.

    With the number 1, there alone, I only ask you to print the line by default, which by now will have already been modified or will be any other that does not meet the conditions of the regex.

    update

    Since I decided to do it with sed, this would be the answer:

    $ sed '/<name>test-plugin<\/name>/{n;s/3.0.0-SNAPSHOT/1.0.0-SNAPSHOT/}' archivo.xml

    What we do is that the name of the plugin will cause whatever is between the braces {} to be executed.

    And what is between the curly braces is:

    • "n", which adds the following line to the pattern space
    • "s/pattern/replacement/", which will replace the pattern with its replacement.

    The truth was not as much pain as I thought it would be.

    • 3

  2. I tried to help you using regexbut it can be quite complex and difficult to maintain since it would not be as flexible to changes (something you have to take into account), even awkthough it is much more powerful than sed, what you can do is generate a script in python so that it does what you need. I leave you the code (personally I prefer the 2 version of this example):

    1st version

    import re 
f = open("/ruta/a_su_archivo/archivo.xml", "r")
lines = f.readlines()
f.close()
plugin_name = "test-plugin"
new_version_of_plugin = '2.5.3'

for i in range(len(lines)):
    if re.search(f"<name>{plugin_name}<\/name>", lines[i]):
        lines[i+1] = re.sub("<version>.*</version>",f"<version>{new_version_of_plugin}</version>", lines[i+1])
        print(lines[i+1])
print(lines)

f = open("/ruta/a_su_archivo/archivo.xml", "w+")
f.writelines(lines)
f.close()

    you can save it with any name, for this example save it as nombre_del_script.pyand call it like this:

    python3 nombre_del_script.py

    You can also receive the file path and version by console, in that case I would change the code to:

    2nd version

    import re 
import sys

file = sys.argv[0]
plugin_name = sys.argv[1]
new_version_of_plugin = sys.argv[2]

f = open(file, "r")
lines = f.readlines()
f.close()

for i in range(len(lines)):
    if re.search(f"<name>{plugin_name}<\/name>", lines[i]):
        lines[i+1] = re.sub("<version>.*</version>",f"<version>{new_version_of_plugin}</version>", lines[i+1])
        print(lines[i+1])
print(lines)

f = open(file, "w+")
f.writelines(lines)
f.close()

    and calls it like this (obviously changes to the corresponding values):

    python3 nombre_del_script.py ruta_al_archivo.xml nombre_del plug_in version_que_quiere_colocar
    • 2