I have a pretty simple but huge xml (more than 50.000 lines). I want to extract just one repeated node as table (ring). Take into account this is an example. The real file has more than 300 elements inside "ring" node, so I try to avoid to write command for extraction element by element if that is possible.
Here a little example
<xml>
<root>
<ration>
<ring>
<id value="1"/>
<date value="2021-01-01"/>
<price value="435"/>
</ring>
</ration>
</root>
<root>
<ration>
<ring>
<id value="14"/>
<date value="2021-02-01"/>
<price value="745"/>
</ring>
</ration>
</root>
</xml>
what I want to achieve is: this xmlstarlet command is working and I get the proper result but now think on doing it 300 times (for every element I have to insert the call into the concat part). I need a more simple approach.
xmlstarlet sel -T -t -m '//root/ration/ring' -v "concat(id/@value,',',date/@value,',',price/@value)" -n file.xml
id, date, price
1, 2021-01-01, 435
14, 2021-02-01, 745
It's been a while since you asked. Nevertheless...
xmlstarlet sel -T \
-t -m '//root[1]/ration/ring/*[boolean(@value)]' \
-v 'substring(", ", 1, 2*(position() != 1))' -v 'local-name()' -b -nl \
-t -m '//root' -m 'ration/ring/*[boolean(@value)]' \
-v 'substring(", ", 1, 2*(position() != 1))' -v '@value' -b -nl \
file.xml
Output:
id, date, price
1, 2021-01-01, 435
14, 2021-02-01, 745
Template #1 emits the header, #2 the data. -m expressions #1 and #3 match child elements of ring having a @value attribute. Using -m and -b together does the trick. If @value is replaced with @* all attributes are selected; with a predicate such as @*[name()="value" or name()="otherval"] selection can be fine-tuned.
Separation of values takes more work in XSLT 1.0 than in 2.0: in this case a substring function call returns an empty string for the first value where position() != 1 is false (converted to number 0), and a 2-character string (comma and space) for the following where it is true (1).
By the way, tag / PI names beginning with xml in any lower/upper combination are reserved.