Shell Programming and Scripting

Hello all,

This is a bit simple but I cannot do it! I have a big table of values between 0 and 1. some cells have NA instead of having any value.

I want to exchange all values with "1" and write "0" for all "NA" .

for changing "NA" to zero I used sed:

Code:

sed -i 's/NA/0/g' input.txt

But I have no clue for changing all numbers to 1.

I appreciate if you can help me with this.

---------- Post updated at 06:54 PM ---------- Previous update was at 06:47 PM ----------

I have managed to do it with this:

Code:

sed -e 's/[0-9]*\.[0-9]*/1/g' -e 's/NA/0/g' in.txt > out.txt

I hope it's correct!

Well, it will also turn . into 0, so in place of * you might want to say \{1,9\}, but for well behaved input, it is ok.

you're right DGPickett! I haven't thought about it! Actually my input is rather big so I have no clue if it's well behaved! Gonna do what you recommended! Thanks!

Yes, I often wish for the meta-version '+' that a few regex flavors have, like * but for one or more.

Limiting length correctly is very important since by default regex is aggressive (greedy -- biggest string that will fit). I get a lot of mileage out of '[^sep]' in place of '.' Some pattern matching tools let you choose lazy or aggressive, like the ksh ${name#pattern} is lazy and ${name##pattern} is aggressive (# is pound on the nose, % is take your percentage in the end).