A log file is of below format

Amazon Interview Question for Applications Developers

0

of 0 votes

6
Answers
A log file is of below format
log.txt
=====================================
TIMESTAMP:MM-DD-YYYY
REQUEST:______________________________
RESPONSE:____________________________
SUCCESS:200:________________________
=====================================
TIMESTAMP:MM-DD-YYYY
REQUEST:______________________________
RESPONSE:____________________________
ERROR:400:________________________
=====================================
TIMESTAMP:MM-DD-YYYY
REQUEST:______________________________
RESPONSE:____________________________
ERROR:=500:________________________
=====================================
TIMESTAMP:MM-DD-YYYY
REQUEST:______________________________
RESPONSE:____________________________
ERROR:400:________________________
=====================================
TIMESTAMP:MM-DD-YYYY
REQUEST:______________________________
RESPONSE:____________________________
ERROR:401:________________________

I want the count of error between two given dates.
```
grep '0[6789]:\| 1[01234567]' | ""grep -c "ERROR" log.txt      

timestamp = grep "TIMESTAMP"
awk '$timestamawk>=from&&$timestamp<=to' from="2007/03/20 15:13" to="2007/08/19 14:31" log.txt | grep "ERROR" | wc -l

timestamp = grep "TIMESTAMP"
echo timestamp | awk -F'[- ]' '$2 >= 25 && $3 <= 04 { print }' | grep -c "ERROR"
```
- anushvenki May 12, 2013 in India | Report Duplicate | Flag | PURGE
Amazon Applications Developer Unix

Email me when people comment.

An error occurred in subscribing you.

Country: India
Interview Type: Phone Interview

More Questions from This Interview

Email me when people comment.

An error occurred in subscribing you.

Comment hidden because of low score. Click to expand.

of 0 vote

sed -n '/STARTDATE/,/ENDDATE/p' log.txt | grep -c ERROR

- Daya May 14, 2013 | Flag Reply

Comment hidden because of low score. Click to expand.

of 0 votes

This would be off by 1 (1 less):
i.e:

07-01-1992
ERROR
07-02-1992
ERROR

Now this will only count the first error if we take it from 07-01-1992 to 07-02-1992, not only that, it will fail if there is no error that start / end date.

- Mo Lam May 15, 2013 | Flag

Comment hidden because of low score. Click to expand.

of 0 votes

It will work only if you take start date and end date as exact given in the file.

- lakshitarora26 June 05, 2013 | Flag

Comment hidden because of low score. Click to expand.

of 2 vote

Use date as the key. Use a cumulative array A that counts the number of errors seen up to (and including) a given date.

Given dates di and dj where di<=dj, the number of errors between those dates is simply
: A[dj]-A[di-1]

- Murali Mohan May 18, 2013 | Flag Reply

Comment hidden because of low score. Click to expand.

of 2 vote

If we want to maintain a dynamic set of records, a still better approach would be to use a balanced BST. The BST would have date as the 'key' field and should be augmented with a 'errorCount' field that maintains the count of errors seen on a given date.

1. As the input file is scanned, insert/update errorCount field of the appropriate key in the BST
2. To find the errorCount in a given range, use a modified version of rank() operation that adds up the errorCount fields is the subtrees.

CareerCup

Amazon Interview Question for Applications Developers

Books

Videos

Resume Review

Mock Interviews