Finding element of file in KSH

Question

please help me figure out what I am doing wrong.

I have this file called "res.xml"

<sequence type="a"> <transaction> <branchSerial>1</branchSerial> <postingDate>2021-08-02</postingDate> <postingTime>2021-08-06 19:42:49 UTC</postingTime> <step type="a"> <record type="c" label="deviceRecord"> <operation option="V">View</operation> <tableName>DEVICE</tableName> <tableDescription>Device</tableDescription> <rowDescription>[email protected]</rowDescription> </record> </step> </transaction> </sequence> 

I am trying to pull the postingDate and put into a variable. My script:

#!/bin/ksh dbDATE=$(sed -n -e "s/<postingDate>\([0-9]*\)<\/postingDate>/\1/p" res.xml) echo "current DB date: $dbDATE" 

when I run it, I get nothing, just blank.

It is very odd because I have used the same logic elsewhere without problems. Can anybody see what I am missing???

Thank you so much for any help you can provide

Answer 1

The main issue with your sed approach is that you do not allow the date to contain dashes.

---

To extract the data from the file that you show, you would use an XML-aware parser for the command line.

One example of such a parser is xmlstarlet, which you would use either like this:

dbDATE=$( xmlstarlet sel -t -v '/sequence/transaction/postingDate' res.xml ) 

or, if the node you're looking for the value of is the only such node,

dbDATE=$( xmlstarlet sel -t -v '//postingDate' res.xml ) 

Another XML-aware parser is xq, an XML parser wrapper around the JSON parser jq:

dbDATE=$( xq -r '.sequence.transaction.postingDate' res.xml ) 

---

All of the above assumes that the sequence node contains only a singletransaction node. Do we want to support multiple transactions (the word "sequence" hints that this may contain a list of transactions), you will also need to decide whether you want to pick the first transaction or specific transaction given some other condition.

Assuming we want the transaction with a branchSerial value of 1. With xmlstarlet:

dbDATE=$( xmlstarlet sel -t -v '/sequence/transaction[branchSerial=1]/postingDate' res.xml ) 

With xq:

dbDATE=$( xq -r '.sequence.transaction[] | select(.branchSerial == "1").postingDate' res.xml ) 

Answer 2

Lets say you set the .xml output to a variable, you can do the following:

xmlData=' <sequence type="a"> <transaction> <branchSerial>1</branchSerial> <postingDate>2021-08-02</postingDate> <postingTime>2021-08-06 19:42:49 UTC</postingTime> <step type="a"> <record type="c" label="deviceRecord"> <operation option="V">View</operation> <tableName>DEVICE</tableName> <tableDescription>Device</tableDescription> <rowDescription>[email protected]</rowDescription> </record> </step> </transaction> </sequence> ' date=$(echo "$xmlData" | grep "postingDate" | tr '>' " " | tr '<' " " | awk '{print $2}') 

---

Since you said you are pulling it in from a file though, you can even do this:

date=$(cat res.xml | grep "postingDate" | tr '>' " " | tr '<' " " | awk '{print $2}') 

Answer 3

works fine with below command

dbdate=$(awk -F "[<>]" '/postingDate/{print $3}' res.xml) echo -e "current DB date: $dbdate" current DB date: 2021-08-02