in need of sed/awk/cut help! thanks!

[djcham]

hello,

I have a very long line, it's a really long line and i want to parse names out of them. Imagine a line like this:

hello my name is = Jonathan<miscellaneous text>hello my name is = Michael<miscellaneous text>hello my name is = Eric

I want to get all the names from there and I don't know the length but I know that following every name will be a "hello" word. Any idea how i can get all these names parsed out into a file like this:

Jonathan
Michael
Eric

Thanks in advance!!!

[smolloy]

Quick guess

awk -F " hello " '{print $NF}' yourfile.txt

I might have got the syntax wrong, but this changes the field separator from a space to " hello " (note the spaces on either side). It then prints out the last word from each "field" -- which should be the names you're looking for.

Apologies if I got the syntax of the command wrong, but something like this should work.

[smolloy]

Sorry, I was wrong. Try this

awk '{print $NF}' RS="hello" yourfile.txt

This changes the record separator, not the field separator, and prints the last field from every record (which is the correct thing to do). In this case the last field before every record separator ("hello") will be a name.

[djcham]

this sort of worked. The problem is I have miscellaneous text in the middle between the hello. so for instance, like

hello my name is = Jonathan how are you hello my name is = Michael what are you doing today? hello my name is = Eric What's up

So there is a space after the name but I won't know exactly how long the miscellenous text will be. is there anyway to AWK out everything between the "=" sign and then space (where the miscellaneous text begins?) thanks again.

[radoulov]

Use perl:

perl -lne'print $1 while /=\s*(\S+)/g' file

[vsemaska]

Problem is you have mutiple occurances on a line. The only way I see how to do this in awk is writing an awk program as follows, let's call in names.awk

{lne=$0; pos=1;
while (pos != 0)
{ pos=index(lne,"hello my name is = ");
if (pos == 0) next;
pos=pos+19;
lne=substr(lne,pos);
$0=lne
print $1 }
}

This will search the line for "hello my name is = ", remove everything before the actual name, and print the name. It repeats this until it can't find another "hello my name is = ".

The command would be:

awk -f names.awk yourfile.txt

[radoulov]

Or:

awk 'NR>1&&$0=$1' RS=\= file

[Steve Yau]

The following works even if your miscellaneous text contains "=":

awk '$0=$1' RS="hello my name is =" <your file>

[radoulov]

Only with GNU Awk though ...

[Steve Yau]

Yes, this will only work with gawk which allows RS to be a regular expression instead of just single character.
Since djcham already mentioned smolloy's suggestion a sort of worked, I assumed he is using a version Linux with gawk as the default awk.