Skip to first unread message
Dilawar Singh
Jan 19, 2014, 1:28:44 AM1/19/14
to wncc...@googlegroups.com
A good way to learn basic *nix is by reading its `info` pages. Command `info` (prefer pinfo if installed) in terminal opens them. Here is some tips on searching files on unix.
If you know the name of the file, searching is as easy as the command `locate filename`.
This is very helpful when you wonder if a library or header file is available on your system or not. You should update your database once in a while by using `sudo updatedb`
command. You can automate it, for example if you want to update your database everyday at 2pm, following would do the job
sudo crontab -e
This command will open a file, append this line,
0,0 14 * * * updatedb
Now you can search quickly for a filename using locate command. For example, if I want to check if a particular boost header is installed on my machine (or where it is), all I need to do is, `locate boost`. This usually prints a lot of names, we can "pipe" its output to `grep`. Command grep is a powerful command and it will filter those files which does not match with grep input.
locate boost | grep graph
(i.e. search for boost but drop all names that do not contain the word graph).
====
Searching files which contain some words in terminal can be tricky. People often combine `find` with `xargs`.
Dilawar
EE, IITB
If you know the name of the file, searching is as easy as the command `locate filename`.
This is very helpful when you wonder if a library or header file is available on your system or not. You should update your database once in a while by using `sudo updatedb`
command. You can automate it, for example if you want to update your database everyday at 2pm, following would do the job
sudo crontab -e
This command will open a file, append this line,
0,0 14 * * * updatedb
Now you can search quickly for a filename using locate command. For example, if I want to check if a particular boost header is installed on my machine (or where it is), all I need to do is, `locate boost`. This usually prints a lot of names, we can "pipe" its output to `grep`. Command grep is a powerful command and it will filter those files which does not match with grep input.
locate boost | grep graph
(i.e. search for boost but drop all names that do not contain the word graph).
====
Searching files which contain some words in terminal can be tricky. People often combine `find` with `xargs`.
Dilawar
EE, IITB
Dilawar Singh
Jan 19, 2014, 1:36:47 AM1/19/14
to wncc...@googlegroups.com
Sorry auto markdown to html script did not work as expected. Here is them complete post.
Locating file using locate
If you know the name of the file (or a pattern of it), searching is as easy as the command `locate filename`. One has to install `mlocate` or equivalent program (on most of system it is already install). This is very helpful when you want to figure out if a library or header file is available on your system. You should update your database once in a while by using `sudo updatedb` command. You can automate it using crontab.
For example if you want to update your database everyday at 1400Hrs, following would do the job
Locating file using locate
If you know the name of the file (or a pattern of it), searching is as easy as the command `locate filename`. One has to install `mlocate` or equivalent program (on most of system it is already install). This is very helpful when you want to figure out if a library or header file is available on your system. You should update your database once in a while by using `sudo updatedb` command. You can automate it using crontab.
For example if you want to update your database everyday at 1400Hrs, following would do the job
sudo crontab -e
This command will open a file, append this line,
0,0 14 * * * updatedb
Now you can search quickly for a filename using locate command in updated
database. If database is not up-to-date then this command might not give you the
location of file.
# Using grep
Let's say I want to check a particular boost header is installed on my machine. If I say `locate boost`. This usually prints a lot of names, we can "pipe" its output to `grep`. Command grep is a powerful command. It filters those files which does not match with grep input.
(i.e. search for boost but drop all names that do not contain the word graph). Try it with without grep and then with grep.
# Finding files
On server machines where the admin may or may not have installed/updated locate database, one can use the powerful `find` command. Command `info find` or `man find` will open up its info and manual page in terminal. This command can take many arguments and they are not always easy to remember. It is always useful to create small scripts for specific purposes. Store those scripts on some online repositories so that they can be downloaded easily later anytime. I keep mine on github (http://github.com/dilawar/Scripts).
A typical example of find: I want to search a header file which I want to include in my C program. I am planning to search it in `/usr/include` folder, if not found I'll try to search in `/usr` directory.
One can also pass `-type f` or `-type d` to search only files or directory respectively. One often uses find command with `xargs` command. Files searched by `find` passed to another command. One example, let's say I want to delete all those files which are updated in previous 10 minutes.
This command will print name of files inside directory `/home/dilawar/.config` which are updated in last 10 minutes. We can pipe the names of these files to some other command e.g. `rm`.
Each line of first command is copied into variable file (-I file) and passed onto the `rm -f file` command which removes the file. (See what -print0 options does).
# Using find with vim
[This script](http://github.com/dilawar/Scripts/v) uses find and git to search for a file, creates its backup and open the file in vim. The default backup directory is `~/.backup` (it is a local git repository). Each change is committed to this directory. This is sometimes useful if you are working outside a version control system. Let's say I have source code in a directory `A` and I know there should be a file which has word `srcinfo` in its name. I do the following v *srcinfo* This will search for files recursively and opens 3 best matches in separate tabs. One can modify it (change vim to emacs or nano or whatever).
# References
- Use `info` command to read the info pages. They are great.
- Checkout The Linux Documentation Project (TLDP) where great many tutorials are written. Various stackexchange sites are great places also for specific question.
- Dilawar
# Using grep
Let's say I want to check a particular boost header is installed on my machine. If I say `locate boost`. This usually prints a lot of names, we can "pipe" its output to `grep`. Command grep is a powerful command. It filters those files which does not match with grep input.
locate boost | grep graph(i.e. search for boost but drop all names that do not contain the word graph). Try it with without grep and then with grep.
# Finding files
On server machines where the admin may or may not have installed/updated locate database, one can use the powerful `find` command. Command `info find` or `man find` will open up its info and manual page in terminal. This command can take many arguments and they are not always easy to remember. It is always useful to create small scripts for specific purposes. Store those scripts on some online repositories so that they can be downloaded easily later anytime. I keep mine on github (http://github.com/dilawar/Scripts).
A typical example of find: I want to search a header file which I want to include in my C program. I am planning to search it in `/usr/include` folder, if not found I'll try to search in `/usr` directory.
find /usr/local -name "*graph.h"
One can also pass `-type f` or `-type d` to search only files or directory respectively. One often uses find command with `xargs` command. Files searched by `find` passed to another command. One example, let's say I want to delete all those files which are updated in previous 10 minutes.
find /home/dilawar/.config -mmin 10
This command will print name of files inside directory `/home/dilawar/.config` which are updated in last 10 minutes. We can pipe the names of these files to some other command e.g. `rm`.
find /home/dilawar/.config -mmin 10 | xargs -I file rm -f file
Each line of first command is copied into variable file (-I file) and passed onto the `rm -f file` command which removes the file. (See what -print0 options does).
# Using find with vim
[This script](http://github.com/dilawar/Scripts/v) uses find and git to search for a file, creates its backup and open the file in vim. The default backup directory is `~/.backup` (it is a local git repository). Each change is committed to this directory. This is sometimes useful if you are working outside a version control system. Let's say I have source code in a directory `A` and I know there should be a file which has word `srcinfo` in its name. I do the following v *srcinfo* This will search for files recursively and opens 3 best matches in separate tabs. One can modify it (change vim to emacs or nano or whatever).
# References
- Use `info` command to read the info pages. They are great.
- Checkout The Linux Documentation Project (TLDP) where great many tutorials are written. Various stackexchange sites are great places also for specific question.
- Dilawar
Reply all
Reply to author
Forward
0 new messages