How to get remote file size from a shell script?

Is there a way to get the size of a remote file like

http://api.twitter.com/1/statuses/public_timeline.json

in shell script?

Solution 1:

You can download the file and get its size. But we can do better.

Use curl to get only the response header using the -I option.

In the response header look for Content-Length: which will be followed by the size of the file in bytes.

$ URL="http://api.twitter.com/1/statuses/public_timeline.json"
$ curl -sI $URL | grep -i Content-Length
Content-Length: 134

To get the size use a filter to extract the numeric part from the output above:

$ curl -sI $URL | grep -i Content-Length | awk '{print $2}'
134

Solution 2:

Two caveats to the other answers:

1. Some servers don't return the correct Content-Length for a HEAD request, so you might need to do the full download.
2. You'll likely get an unrealistically large response (compared to a modern browser) unless you specify gzip/deflate headers.

Also, you can do this without grep/awk or piping:

curl 'http://api.twitter.com/1/statuses/public_timeline.json' --location --silent --write-out 'size_download=%{size_download}\n' --output /dev/null

And the same request with compression:

curl 'http://api.twitter.com/1/statuses/public_timeline.json' --location --silent  -H 'Accept-Encoding: gzip,deflate' --write-out 'size_download=%{size_download}\n' --output /dev/null

Solution 3:

Similar to codaddict's answer, but without the call to grep:

curl -sI http://api.twitter.com/1/statuses/public_timeline.json | awk '/Content-Length/ { print $2 }'

Solution 4:

The preceding answers won't work when there are redirections. For example, if one wants the size of the debian iso DVD, he must use the --location option, otherwise, the reported size may be that of the 302 Moved Temporarily answer body, not that of the real file.
Suppose you have the following url:

$ url=http://cdimage.debian.org/debian-cd/8.1.0/amd64/iso-dvd/debian-8.1.0-amd64-DVD-1.iso

With curl, you could obtain:

$ curl --head --location ${url}
HTTP/1.0 302 Moved Temporarily
...
Content-Type: text/html; charset=iso-8859-1
...

HTTP/1.0 200 OK
...
Content-Length: 3994091520
...
Content-Type: application/x-iso9660-image
...

That's why I prefer using HEAD, which is an alias to the lwp-request command from the libwww-perl package (on debian). Another advantages it has is that it strips the extra \r characters, which eases subsequent string processing.

So to retrieve the size of the debian iso DVD, one could do for example:

$ size=$(HEAD ${url})
$ size=${size##*Content-Length: }
$ size=${size%%[[:space:]]*}

Please note that:

• this method will require launching only one process
• it will work only with bash, because of the special expansion syntax used

For other shells, you may have to resort to sed, awk, grep et al..

Solution 5:

I think the easiest way to do this would be to:

1. use cURL to run in silent mode -s,

2. pull only the headers -I (so as to avoid downloading the whole file)

3. then do a case insensitive grep -i

4. and return the second arg using awk $2.

5. output is returned as bytes

Examples:

curl -sI http://api.twitter.com/1/statuses/public_timeline.json | grep -i content-length | awk '{print $2}'

//output: 52

or

curl -sI https://code.jquery.com/jquery-3.1.1.min.js | grep -i content-length | awk '{print $2}'

//output: 86709

or

curl -sI http://download.thinkbroadband.com/1GB.zip | grep -i content-length | awk '{print $2}'

//output: 1073741824

Show as Kilobytes/Megabytes

If you would like to show the size in Kilobytes then change the awk to:

awk '{print $2/1024}'

or Megabytes

awk '{print $2/1024/1024}'