In CSV file, Based on match pattern,fetch the other value matched pattern

awk -F ',' '
BEGIN { OFS = FS }
{
    a = t = "N/A"

    for (i = 1; i < NF; ++i)
        if ($i == "AMOUNT")
            a = $(i + 1)
        else if ($i == "TXN")
            t =$(i + 1)

    $0 = ""

    $1 = "AMOUNT"
    $2 = a
    $3 = "/AMOUNT"

    $4 = "TXN"
    $5 = t
    $6 = "/TXN"

    print
}' file

This looks for the strings AMOUNT and TXN within the comma-delimited fields of each line. When AMOUNT is found, a is assigned the data of the next field. In a similar way, t is assigned the TXN value. If either of the values can't be found, it will be the string N/A.

The output is then generated.

A shorter version:

awk -F ',' '
{
    a = t = "N/A"

    for (i = 1; i < NF; ++i)
        if ($i == "AMOUNT")
            a = $(i + 1)
        else if ($i == "TXN")
            t =$(i + 1)

    printf "AMOUNT,%s,/AMOUNT,TXN,%s,/TXN\n", a, t
}' file

Only the outputting bit is different. Instead of outputting a record, we output a string produced by printf.