SED command contained in nested loops

I have a scenario wherein a file contains xml tags. There is a tag like <abcd .... param="paramvalue"> under which I have another tag named <efgh="value"> under which another tag <hijk> is present. I need to count the no. of <hijk> tags present in each <abcd> tag and I have to display the value of param in <abcd> tag and value of <efgh> tag.

For example,

FILENAME.xml

<abcd...... xyz="value1">
      <efgh="value2">
          <hijk>
            ....
          </hijk>
           ....
      </efgh>
      <efgh="value3">
         ....
     </efgh>
       ....
</abcd>

<abcd...... xyz="value4">

      .......

</abcd>

    ......

I used

sed -n '/xyz=*/p' FILENAME.xml | cut -d =  -f 4 | echo "xyz:` cut -d " " -f 1`" && sed -n '/efgh=*/p' FILENAME.xml | cut -d =  -f 2 | echo "efgh:`cut -d " " -f 1`" && grep -c '`<hijk>`' FILENAME.xml

But I got the total count of <hijk> tags and got a separate list of xyz values and a separate list of efgh values.

I need the solution in this format,

xyz="value1"

  efgh="value2"

   no. of hijk tags

....

xyz="valueN"

   efgh="valueN"

   no. of hijk tags

.... 

and so on.

And I thought I could use 3 inner loops along with sed command to iterate. Am I correct? Even if I am, I need help doing so. Please help.