How do I replace multiple fields in multiple XML files?

I have around 4000 XML files and I need to replace the value of both <filename> and <path> fields. I need to replace those fields dynamically. e.g. images0001.xml should have images0001 inside the two fields, images0002.xml should have images0002 inside the two fields, etc.

I've already used this command to rename the files sequentially:

rename 's/.+/our $i; sprintf("images%04d.jpg", 1+$i++)/e' *

And I also used this command to delete the .jpg extension that was in the two fields I'm trying to change:

sed -i 's/.jpg//g' Annotations/*

Here is the current state of the contents of the XML files:

<annotation>
    <folder></folder>
    <filename>1608644703_2.rf.fa179c1e6c47d72d668ac3d83c7f79d1</filename>
    <path>1608644703_2.rf.fa179c1e6c47d72d668ac3d83c7f79d1</path>
    <source>
        <database>roboflow.ai</database>
    </source>
    <size>
        <width>416</width>
        <height>416</height>
        <depth>3</depth>
    </size>
    <segmented>0</segmented>
    <object>
        <name>megot</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <occluded>0</occluded>
        <bndbox>
            <xmin>129</xmin>
            <xmax>292</xmax>
            <ymin>145</ymin>
            <ymax>351</ymax>
        </bndbox>
    </object>
</annotation>

And here is how I need the files to be changed:

<annotation>
    <folder></folder>
    <filename>images0001</filename>
    <path>images0001</path>
    <source>
        <database>roboflow.ai</database>
    </source>
    <size>
        <width>416</width>
        <height>416</height>
        <depth>3</depth>
    </size>
    <segmented>0</segmented>
    <object>
        <name>megot</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <occluded>0</occluded>
        <bndbox>
            <xmin>129</xmin>
            <xmax>292</xmax>
            <ymin>145</ymin>
            <ymax>351</ymax>
        </bndbox>
    </object>
</annotation>

I'm looking for a way to do this in command line, but I can't figure out a solution after searching for a while!

Any help will be appreciate. Thanks in advance!

One way to do it with sed is by running the following command:

for f in *.xml; do sed "s|\(<filename>\).*\(</filename>\)|\1${f%.*}\2|; s|\(<path>\).*\(</path>\)|\1${f%.*}\2|" "$f"; done

• for f in *.xml; do ... ; done is a basic for loop for .xml files in your current directory. Each file found is stored in the f variable.

• sed "s|\(<filename>\).*\(</filename>\)|\1${f%.*}\2|; s|\(<path>\).*\(</path>\)|\1${f%.*}\2|" "$f" is the command that is run for each file found. The command does two similar replacements, one for the <filename> field and one for the <path> field:

  • s|\(<filename>\).*\(</filename>\)|\1${f%.*}\2| matches the text \(<filename>\).*\(</filename>\) using regex (.* means match any character between <filename> and </filename>). The \( and \) are not matched but are used to capture the <path> and </filename> texts which are then used as replacements. Then the text is replaced by the first capture \1 (<filename>), the filename without the extension ${f%.*}, and the second capture \2 (</filename>).

  • Similarly for path, using it instead of filename.

After making sure that the above command works as intended by using it in a copied portion of your files, add the -i flag right after sed to change the files in place.