CODEWITHSUNDEEP
bashfindunix-timestampgnu-findutils
Dan Parker
Dan Parker

Reputation: 3

Find files in directory modified at various times

Bash newbie here. The task is to find all files in a (rather large) directory whose modification date falls between various pairs of times. These times are specified as Unix timestamps in a CSV file:

1483743420,1483747020
1484348640,1484352240
1484953920,1484957520
1485559200,1485562800
1486164480,1486168080
1486769820,1486773420

My early thought was to use find and awk:

find "$PROJECT_DIR" -type f \
-newermt "$(awk -F "","" '{print $1}' "$DATE_CSV")" -not \
-newermt "$(awk -F "","" '{print $2}' "$DATE_CSV")"

find doesn't seem to recognize this date format. More pressingly, this code only sends the first date to find, and the rest are printed:

$ bash datefind.sh
find: Can't parse date/time: 1483743420
1484348640
1484953920
1485559200
1486164480
1486769820

Is there a way to do this with a single find command? Or should I be trying something else?

Upvotes: 0

Views: 82

Answers (1)

Cyrus
Cyrus

Reputation: 88583

With GNU find:

PROJECT_DIR="/tmp"
DATE_CSV="file"
find "$PROJECT_DIR" \( -type f \) -a \( $(awk -F ',' 'NR==1 { print "-newermt @" $1 " -not -newermt @" $2 }; NR!=1 { print "-o -newermt @" $1 " -not -newermt @" $2 }' "$DATE_CSV") \)

From man find:

-newermt: Time specifications are interpreted as for the argument to the -d option of GNU date.

-o: logical OR

-a: logical AND

What awk does:

DATE_CSV="file"
awk -F ',' 'NR==1 { print "-newermt @" $1 " -not -newermt @" $2 }; NR!=1 { print "-o -newermt @" $1 " -not -newermt @" $2 }' "$DATE_CSV"

Output:

-newermt @1483743420 -not -newermt @1483747020
-o -newermt @1484348640 -not -newermt @1484352240
-o -newermt @1484953920 -not -newermt @1484957520
-o -newermt @1485559200 -not -newermt @1485562800
-o -newermt @1486164480 -not -newermt @1486168080
-o -newermt @1486769820 -not -newermt @1486773420

Upvotes: 1

Related Questions