If we take your two commands (modified to create file3 instead of append to it):

Code:

sed -n '/phone/{s/.*<phone>\(.*\)<\/phone>.*/\1/;p;}' file1 > file2  
awk -F '[<>]' '{n = split($3,array," "); print array[n] > "file3" }' file2

you get what you want in file3. If you change the 2nd command to:

Code:

awk -F '[<>]' '{print $3}' file2 > file4

you get the same output in file4. And, if you change the 1st command to:

Code:

sed -n '/phone/{s/.*<number>\(.*\)<\/number>.*/\1/;p;}' file1 > file5

you get the same output in file5 with one command instead of two. And, if you prefer to use awk instead of sed, with your sample input, the following:

Code:

awk -F '[<>]' '/phone/{print $5}' file1 > file6

produces the same output in file6. But, if there could be other tags between the phone tags, the following would be safer:

Code:

awk -F '<number>|</number>' '/phone/{print $2}' file1 > file7

still producing the same output in file7.

Of course, the 1st two of the above work only if the number tags are on the same line as the starting and closing phone tags; the others work as long as the starting and closing number tags are on the same line as a starting or closing phone tag.

None of the above work if the starting and closing number tags are not on the same line. And, none of the above will find all of the numbers you want if there is more than one set of number tags on a single line.