Tips and Tricks

I always find it difficult to digest the filesize from the ls -al command. For instance, after ls -al, the output give me filesize in bytes.

Gosh, then I have to start calculating it by taking last 4 digits, slowly count upwards like 1K, 10K, 100K, 1MB, 10MB, 100MB and so on so forth.

For instance:

this output:
-rw-r--r-- 1 walrus dba 137207094 Jul 8 23:12 config.2008032519.s

137207094 is how much?

going with my method of counting upwards, it gives me 137MB roughly.

Is it correct? WRONG. Hell wrong

The above is bits only. Bear in mind, 1 KB = 1024 bits, 1 MB = 1024 KB and so on so forth

[ Source ]
1 bit = a 1 or 0 (b)
4 bits = 1 nybble (?)
8 bits = 1 byte (B)
1024 bytes = 1 Kilobyte (KB)
1024 Kilobytes = 1 Megabyte (MB)
1024 Megabytes = 1 Gigabyte (GB)
1024 Gigabytes = 1 Terabyte (TB)

The correct calculation is 137207094 / 1024 (bits) / 1024 (KB) = 130.8 MB

Starting in Solaris 10, we have a new option in ls command.

: /u01/apps/WatchMark/FlexPM/classic/vendor/Lucent/ECP/ftpIN/in>uname -a
SunOS lxserver 5.10 Generic_118833-36 sun4u sparc SUNW,Sun-Fire-V445

ANCIENT WAY:
ls -al
total 270388
drwxr-xr-x 2 walrus dba 1024 Jul 8 23:14 .
drwxr-xr-x 11 walrus dba 512 Jun 17 01:49 ..
-rw-r--r-- 1 walrus dba 137207094 Jul 8 23:12 config.2008032519.s
-rw-r--r-- 1 walrus dba 451989 Jul 8 23:12 config.2008032519.split0.bz

cons: hard to read filesize and output distorted

NEW WAY:
ls -alh
total 269060
drwxr-xr-x 2 flexpm dba 1.0K Jul 8 23:12 .
drwxr-xr-x 11 flexpm dba 512 Jun 17 01:49 ..
-rw-r--r-- 1 flexpm dba 131M Jul 8 23:12 config.2008032519.s
-rw-r--r-- 1 flexpm dba 441K Jul 8 23:12 config.2008032519.split0.sm.gz

pros:
- more readable format in terms of file size
- contents are properly aligned.

cons:
- need to type extra 'h' at the end of ls command

How to use regular expression (regex) in grep or vi?

If you need to search for a pattern of "string" or "character" in a file / many files, you will need to use regular expression (regex) either in vi editor or in grep command. But how? See below:

using vi:
- press '/' for searching
- for instance, i wanted to search for CL000, CL001, CL002, ie: the pattern is CLXXX follow by digit.
- syntax:
/^CL[0-9]*

^ - start of pattern
* - zero or more digit

using unix grep command:
- grep CL[0-9]