getting two lines from shell

Question

I use

cmd | egrep 'ID|Value'

andget the following output on my command line:

    ID: 2
Value: 21
    ID: 1
Value: 25
    ID: 1
Value: 22
    ID: 3
Value: 56
    ID: 1
Value: 23
    ID: 2
Value: 50
    ID: 2
Value: 56
    ID: 3
Value: 11
    ID: 3
Value: 26

I would like to filter it by one special ID. But when I use:

egrep 'ID: 1|Value'

It will show the same output with just ID 1, but the Values of all IDs.

I already read about awk, sed and regex, but i can't find anything that would really help me, but I'm also not that deep into that.

Someone can suggest an idea, or tell me that it's the wrong way to approach it and I should do it differently and how?

Output as wanted:

    ID: 1
Value: 25
    ID: 1
Value: 22
    ID: 1
Value: 23

so you want to show the block containing `ID: 1` together with the following line containing `Value: 25`? — fedorqui, Jun 14 '16 at 12:59

fedorqui · Answer 1 · 2016-06-14T13:11:34.340

8

Just tell grep to print the matching line together with the next one:

$ grep -A1 'ID: 1$' file
    ID: 1
Value: 25
    ID: 1
Value: 22
--
    ID: 1
Value: 23

From man grep:

-A NUM, --after-context=NUM

Print NUM lines of trailing context after matching lines. Places a line containing a group separator (--) between contiguous groups of matches. With the -o or --only-matching option, this has no effect and a warning is given.

Note the usage of $ to mark the end of the line. Otherwise, this would match lines line ID: 1323423423.

If you don't like the -- appearing in between matches, get rid of it with:

grep -A1 --no-group-separator 'ID: 1$' file
#        ^^^^^^^^^^^^^^^^^^^^

edited Jun 14 '16 at 13:11

answered Jun 14 '16 at 13:05

fedorqui

275,237
103
548
598

Yep, that should suffice. – Alan Jun 14 '16 at 13:06
2

Note that -A1 means "After", if you wanted the line before you could use -B1 ("Before" – Alan Jun 14 '16 at 13:09
2

@Alan exactly. And if your "after" and "before" match, you can join them using `-C`. – fedorqui Jun 14 '16 at 13:09
seems that i didn't read the man as carefully as i thought. It works fine, but my output is coming in blocks now...!? it's like he collects enough input to print it in the screen. any ideas how to get more "realtime" output? – AdrianL Jun 14 '16 at 16:13
@DaPole not sure what you mean here. Are you grepping a continuous stream? – fedorqui Jun 14 '16 at 20:53
Yes, I'm grepping a continious stream of BLE data. – AdrianL Jun 14 '16 at 21:23
@DaPole then you need to use [`--line-buffered`](http://stackoverflow.com/a/7162898/1983854) – fedorqui Jun 14 '16 at 21:35
@fedorqui no effect at all. Stil "buffering" – AdrianL Jun 15 '16 at 08:00

glenn jackman · Answer 2 · 2016-06-14T13:34:26.400

6

For other tags:

awk

awk '/ID: 1$/ {print; getline; print}'

sed, very similar

sed -n '/ID: 1$/ {p;n;p}'

and bash

while read line; do if [[ $line == *"ID: 1" ]]; then echo "$line"; read line; echo "$line"; fi; done

edited Jun 14 '16 at 13:34

answered Jun 14 '16 at 13:30

glenn jackman

238,783
38
220
352

2

and that's the problem with not cutting-and-pasting. – glenn jackman Jun 14 '16 at 13:34

score 5 · Answer 3 · answered Jun 14 '16 at 13:28

5

Another simple script using awk

awk '/ID: 1/{line[NR]; line[NR+1]}; NR in line' file

will give output as:-

    ID: 1
Value: 25
    ID: 1
Value: 22
    ID: 1
Value: 23

NR is a special awk variable which stores the record number. All we do is match a pattern, and get the Nth and N+1th record in the line to print in the file and ensure that the record to be printed is well within the max range in that line (NR in line).

answered Jun 14 '16 at 13:28

Inian

80,270
14
142
161

1

This is first time I have seen this kind of use for `NR`... +1 for creativity. Hopefully I will remember it... – anishsane Jun 14 '16 at 15:17

score 1 · Answer 4 · answered Jun 15 '16 at 11:00

A solution using GNU sed (sometimes available as gsed on BSD systems):

$ sed -n -e '/ID: 1/,+1p' data.in
    ID: 1
Value: 25
    ID: 1
Value: 22
    ID: 1
Value: 23

Explanation: The sed script "/ID: 1/,+1p" applies the p (print) command on all lines of the input in the range /ID: 1/ to +1, i.e every line matching the regular expression "ID: 1" and "one line further along" (the +n address is a GNU extension to the POSIX standard).

The -n flag to sed ensures that nothing other than what the p command prints gets printed ("suppress automatic printing of pattern space").

score 0 · Answer 5 · answered Jan 14 '17 at 13:49

0

This might work for you (GNU sed):

sed '/ID: 1$/!d;n' file

Delete any line that is not ID: 1 and then print it and the next.

answered Jan 14 '17 at 13:49

potong

55,640
6
51
83

getting two lines from shell

5 Answers5