Well, I like databases and thought why don't make a small relational database based on flat files and do joining them using Linux/Unix shell. I tried to invent this small example:

1. Have 2 flat text files, one for Employees (emp.tbl) and the other for Departments (dept.tbl), like the following:

emp.tbl content:

emp_id emp_name dept_id
1      Rami     1
2      Emad     1
3      Mary     2
4      Hassan   1
5      Rasha    2

---

Why google is developing everything (GTalk, Google Desktop, ...) for Microsoft Windows only and not for Linux too, knowing that Linux is being very popular now?!!!

I frequently need to be interactively email me from *NIX server when some output file is generated. I gathered here some used methods:

1. Using mailx for text files:

$ more filename.txt | mailx -s 'subject' email1@company.com, email2@company.com

2. Using mailx for binary files:

$ uuencode filename.bin | mailx -s 'subject' email@company.com

3. Using elm to attach files:

$ echo '[include filename.xls applications/msexcel base64]' > tempfile.txt
$ elm -s 'Subject' email@company.com < tempfile.txt

We compress large files and ftp them to another server regularly. This operation could take about 9hrs. We are using gtar (tar+gzip) to compress the files. I thought we may use bzip2 to decrease the compressed files and save time in ftp, so I do assessment to compare between gzip and bzip2.

Here are the assessment table:

I faced a interesting case: I have a script that generate a line every 10 minutes. This line is appeneded to one single file. You can imagine how big this file will be, if not cleaned regularly. But everytime, when I clean it, I should keep the first and last 2 lines. How can I do this using shell?

It's a piece of cake using sed. Find below the script: