grep for Error and print all the lines containing 2 strings above and below Error

Question

saaa vcahJJ HKak vk     
Import xxx xXXXXX xxxx
aaaa aaaa  aaaa ffffff
hhhhhh hhhhhh hhh hhh  hhhhhh
Error reading readStatus api
aaa hhhh aaa aaaaa
gggggggg ggggg xxxxxxxxxx
uuuu hhhhhhhh fffffffff
query run ends
qidIdih II v iQE Iqe

I want to find the 'Error' string in the file containing above logs and then print all the info available between 2 strings 'Import' and 'ends'. How can I do this using grep/sed Tried this but didn't get much. Note: I dont know how many lines will be before and after. It may vary from above sample I have provided

@tripleee , I am not sure it"s a duplicate. between two marker yes, but there is one more condition here. — ctac_, Nov 21 '17 at 19:51
Thanks for the feedback; removed the duplicate flag albeit hesitantly. Do you really think this needs to exist as a separate question? — tripleee, Nov 21 '17 at 19:57
@tripleee. I don't know exactly. I have'nt read all the aswer in the link but I'm not sure the OP can find a solution for him in the thread. — ctac_, Nov 21 '17 at 20:16

Abhijeetk431 · Answer 1 · 2017-11-21T12:36:06.510

0

Grep's -A 1 option will give you one line after; -B 1 will give you one line before; and -C 1 combines both to give you one line both before and after.

grep -C 1 "Error" <logfile>

As per your requirement, You can use-

sed -n '/Import/,/ends/p' filename

edited Nov 21 '17 at 12:36

answered Nov 21 '17 at 12:10

Abhijeetk431

847
1
8
18

sorry for not being clear in my query, I dont know how many lines will be before and after. It may vary from above sample I have provided – Praveen Nov 21 '17 at 12:13
is there any way to check specifically for 'Import' and 'ends' and grep the contents between these 2 – Praveen Nov 21 '17 at 12:24
Yes.. Edited the answer. Please check – Abhijeetk431 Nov 21 '17 at 12:34

score 0 · Answer 2 · answered Nov 21 '17 at 12:27

0

How about:

$ awk 'BEGIN{RS=ORS="ends\n"} /Error/' file

RS is the input record separator which needs to be ends. ORSgets the same value for output purposes. Also, your example had /^Error/ but Error does not start the record (^).

answered Nov 21 '17 at 12:27

James Brown

36,089
7
43
59

how to do if i have some more lines before 'Import' line – Praveen Nov 21 '17 at 12:50
Please, update the sample data to reflect the real situation. Guessing game is not appreciated here as it's a waste of everyones' time. – James Brown Nov 21 '17 at 12:59

score 0 · Answer 3 · answered Nov 21 '17 at 12:48

0

Here you are:

out=$( sed -n '/^Import/,/end$/p' file )
echo "$out" | grep Error >/dev/null && echo "$out"

This will capture the text between "Import" and "end" and print it only if the extracted text contains "Error".

answered Nov 21 '17 at 12:48

Fabien Bouleau

464
3
11

score 0 · Answer 4 · answered Nov 21 '17 at 21:35

You can try this sed

sed '/^Import/!d;:A;N;/\nImport/{h;s/.*\n//;bA};/ends$/!bA;h;s/\nError//;tB;d;:B;x' infile

Explanation :

sed '
/^Import/!d                 # if a line start with Import
:A
N                           # get an other line
/\nImport/{h;s/.*\n//;bA}   # if the last line start with Import
                            # keep only this last line and return to A
/ends$/!bA                  # If the last line end with ends
h                           # keep all the lines in the hold space
s/\nError//                 # look for a line which start with Error
tB                          # if find jump to B
d                           # not find, delete all and return to the start
:B
x                           # restore all the lines and print
' infile

grep for Error and print all the lines containing 2 strings above and below Error

4 Answers4

Linked