UNIX for Dummies Questions & Answers

Hello all,
I have a test file that has the format:

.....
O
3.694950 -.895050 1.480000
O
5.485050 .895050 1.480000
Ti
-4.590000 4.590000 2.960000
Ti
-2.295000 2.295000 1.480000
O
-5.485050 -5.485050 -4.440000
Ti
.000000 4.590000 2.960000
O
3.694950 -5.485050 -4.440000
....

Basically there is a name/descriptor on one line, followed by the data attached to that name/descriptor on the next line. I'm trying to sort the data so that all the "Ti' lines are together, and all the "O" lines are together. Something like:

Ti
data
Ti
data
Ti
data
...
O
data
O
data
...

I've tried using grep like "grep Ti datafile > out1" and "grep O datafile > out2"
"cat out1 out2 > outnew", but this ignores the data lines. The sort command puts all the data lines together, followed by O lines, followed by Ti lines. I want to keep the data attached to the descriptor, which I can't get sort to do. I can't seem to find any switches in grep or sort that will help me out.

Is there a way to solve my problem with a simple command? I'm sure I could write a python script to solve this, but a command-line solution would be nice.

Thanks a lot.

try this.

https://www.unix.com/unix-for-advanced-and-expert-users/25403-grep-post99821.html#post99821

You could use paste to join the two lines, sort then use tr to split, e.g...

Code:

paste - - < file1 | sort | tr \\t \\n > file2

Hmmm... My grep version (from 'man grep') doesn't seem to support those flags.

Hmmm... could you explain this command? I admit that it leaves me for a bit of a loss. It looks like a lot of piping, and I'm not following it all.

Thanks,
Aaron

that's pretty good, Ygor!

1. 'paste' - joins adjacent lines and separates them with TAB [by default]
2. 'sort' - well.... sorts its input
3. 'tr' - transliterates 'input' chars to the 'output' chars - in our case: TAB to NL (breaking down the lines we've joined in step 1.)