I want to check in linux bash whether a file was created more than x time ago.
let's say the file is called text.txt and the time is 2 hours.
if [ what? ]
then
echo "old enough"
fi
I want to check in linux bash whether a file was created more than x time ago.
let's say the file is called text.txt and the time is 2 hours.
if [ what? ]
then
echo "old enough"
fi
Only for modification time
if test `find "text.txt" -mmin +120`
then
echo old enough
fi
Or, the same in one line:
#!/bin/bash
find text.txt -mmin +120 -exec echo "old enough" \;
You can use -cmin
for change or -amin
for access time. As others pointed I don’t think you can track creation time.
I always liked using date -r /the/file +%s
to find its age.
You can also do touch --date '2015-10-10 9:55' /tmp/file
to get extremely fine-grained time on an arbitrary date/time.
Creation time isn't stored.
What are stored are three timestamps (generally, they can be turned off on certain filesystems or by certain filesystem options):
a "Change" to the file is counted as permission changes, rename etc. While the modification is contents only.
Using the stat
to figure out the last modification date of the file, date
to figure out the current time and a liberal use of bashisms, one can do the test that you want based on the file's last modification time1.
if [ "$(( $(date +"%s") - $(stat -c "%Y" $somefile) ))" -gt "7200" ]; then
echo "$somefile is older then 2 hours"
fi
While the code is a bit less readable then the find
approach, I think its a better approach then running find
to look at a file you already "found". Also, date manipulation is fun ;-)
%Z
instead of %Y
below to get "change time" which may be what you want.[Update]
For mac users, use stat -f "%m" $somefile
instead of the Linux specific syntax above
Although ctime isn't technically the time of creation, it quite often is.
Since ctime it isn't affected by changes to the contents of the file, it's usually only updated when the file is created. And yes - I can hear you all screaming - it's also updated if you change the access permissions or ownership... but generally that's something that's done once, usually at the same time you put the file there.
Personally I always use mtime for everything, and I imagine that is what you want. But anyway... here's a rehash of Guss's "unattractive" bash, in an easy to use function.
#!/bin/bash function age() { local filename=$1 local changed=`stat -c %Y "$filename"` local now=`date +%s` local elapsed let elapsed=now-changed echo $elapsed } file="/" echo The age of $file is $(age "$file") seconds.
The find one is good but I think you can use anotherway, especially if you need to now how many seconds is the file old
date -d "now - $( stat -c "%Y" $filename ) seconds" +%s
using GNU date
Consider the outcome of the tool 'stat':
File: `infolog.txt'
Size: 694 Blocks: 8 IO Block: 4096 regular file
Device: 801h/2049d Inode: 11635578 Links: 1
Access: (0644/-rw-r--r--) Uid: ( 1000/ fdr) Gid: ( 1000/ fdr)
Access: 2009-01-01 22:04:15.000000000 -0800
Modify: 2009-01-01 22:05:05.000000000 -0800
Change: 2009-01-01 22:05:05.000000000 -0800
You can see here the three dates for Access/modify/change. There is no created date. You can only really be sure when the file contents were modified (the "modify" field) or its inode changed (the "change" field).
Examples of when both fields get updated:
"Modify" will be updated if someone concatenated extra information to the end of the file.
"Change" will be updated if someone changed permissions via chmod.