Vim Tips Wiki
Register
(→‎Comments: Use ':sort u')
(example for delimiter sort was wrong, move to comments and explain)
Tag: rollback
(21 intermediate revisions by 11 users not shown)
Line 1: Line 1:
{{review}}
 
 
{{TipImported
 
{{TipImported
 
|id=1166
 
|id=1166
 
|previous=1164
 
|previous=1164
 
|next=1167
 
|next=1167
|created=March 9, 2006
+
|created=2006
 
|complexity=basic
 
|complexity=basic
 
|author=Robert Stovall
 
|author=Robert Stovall
|version=5.7
+
|version=7.0
 
|rating=29/17
 
|rating=29/17
 
|category1=
 
|category1=
 
|category2=
 
|category2=
 
}}
 
}}
  +
Vim has a very powerful built-in sort utility, or it can interface with an external one. In order to keep only unique lines in Vim, you would:
This is just a reminder of how to take advantage of existing Vim capability. I had not needed to sort before and took a few minutes to find a good example.
 
   
Sort lines in Vim using blocks:
 
 
<pre>
 
<pre>
:<range>!sort
+
:{range}sort u
 
</pre>
 
</pre>
   
 
Yes, it's that simple.
 
Yes, it's that simple.
   
  +
You could create a range in advance, such as <code>'a,.</code> (from mark 'a' to the current line) or you could create one on-the-fly using visual selection by pressing ':' in visual mode, after selecting the text you wish to sort, to get a range of <code>'<,'></code> on the command line.
Key strokes:
 
  +
*Place cursor at first line of range to be sorted.
 
  +
If you like using an external sort utility instead, you can do it just as easily. For example, Unix sort, removing duplicate lines:
*Use marker (ma) to mark starting point
 
  +
*Go to last line of range to be sorted
 
 
<pre>
*Issue command from marker (a) to here with <tt>:'a,.!sort</tt>
 
 
:{range}!sort -u
 
</pre>
  +
  +
Many other systems also have an external sort utility, but the options and capabilities will differ. It is probably better to use the built-in Vim sort unless you are looking for a specific feature of the external sort (or using an old Vim without the <code>:sort</code> command).
  +
  +
==Examples==
  +
===Sort in reverse===
  +
<pre>:%sort!</pre>
  +
  +
===Sort, removing duplicate lines===
  +
<pre>:%sort u</pre>
  +
  +
===Sort using the external Unix sort utility, respecting month-name order===
  +
<pre>:%!sort -M</pre>
  +
("respecting month-name order" means January < February < ... < December)
  +
  +
===Numeric sort===
  +
<pre>:sort n</pre>
  +
(this way, 100 doesn't precede 20 in the sort)
  +
  +
===Sort subsections independently, in this example sort numbers between "start" and "end" markers===
  +
<pre>:g/start/+1,/end/-1 sort n</pre>
   
 
==See also==
 
==See also==
Line 33: Line 53:
 
*[[VimTip800|800 Sorting lines in a file based on the number of words in each line]]
 
*[[VimTip800|800 Sorting lines in a file based on the number of words in each line]]
 
*[[VimTip923|923 Sort lines by a specified word number]]
 
*[[VimTip923|923 Sort lines by a specified word number]]
  +
*[[VimTip667|667 Working with CSV files]] sort by CSV field
*[[VimTip1148|1148 Unique sorting]]
 
  +
*[[VimTip128|128 Use Unix command-line tools in Windows]] links to download GNU sort for Windows
  +
*[[VimTip648|648 Uniq - Removing duplicate lines]] techniques to remove duplicate lines
   
 
==References==
 
==References==
Line 39: Line 61:
   
 
==Comments==
 
==Comments==
  +
 
{{Todo}}
 
{{Todo}}
  +
*Probably need some general <code>:sort</code> command info.
Update tip for Vim 7. It is now almost always better to use the <tt>:sort</tt> command that is built into Vim: Use flag <tt>n</tt> to sort numerically or <tt>u</tt> to keep unique lines.
 
  +
*Give examples of numeric sort and using regex sort.
  +
*Clean up my "see also" list. It's useful now for a comprehensive list of related tips, some of which need work. At least should add a note on point of tip.
  +
*If we're going to mention an external sort tool, we may as well include the following with a brief explanation. Vim could do this, but only with a complex regex. Or perhaps better, mention it in [[VimTip374]] or [[VimTip923]] in "see also". <code>-k2</code> sorts on the second field (word by default).
 
<pre>
 
:!sort -k2
 
</pre>
   
 
----
 
----
  +
This misguided snippet was added recently:
Press <tt>V</tt> and highlight the lines you want (instead of marks), then <tt>:!sort</tt>.
 
   
  +
:delimit the column using some char here I have | symbol as delimiter, once did with that you can use below command to sort specific column use -n if u want to sort numeric and its working on some version of vi and not on ubuntu vi :(
----
 
Note that for Vim <7, this uses the external sort program. This has some consequences:
 
*The program must be installed (which it usually is).
 
*Options can be passed to it.
 
   
  +
<pre>/|.*|/ | sort</pre>
which means you can do (for example)
 
<pre>
 
:!sort -n
 
</pre>
 
   
  +
:used to match a patern |.*| used to match words delimited between || and | as piping commend and sort to sort
to sort numerically,
 
   
  +
This is wrong and should never work. Here's what it is actually doing:
<pre>
 
:!sort -k2
 
</pre>
 
   
  +
<code>/|.*|/</code>: jump to the next line that has two '|' characters in it, anywhere
to sort on the second field, etc.
 
   
  +
<code>|</code>: command separator, this lets you start a new command on the current line
----
 
Just to point out that sort is a *nix utility. Windows users will need to get sort.exe from http://www.cygwin.com/ or http://unxutils.sourceforge.net/
 
   
  +
<code>sort</code>: do a default sort of the entire buffer
To sort whole file, and delete duplicates
 
   
  +
Basically this is the equivalent of typing <code>:%sort</code>.
<pre>
 
:%!sort -u
 
</pre>
 
   
  +
Now, what you CAN do, is provide a pattern that the <code>:sort</code> command will skip over and ignore at the start of every line while sorting. For example, to sort based only on text after the last '|' character on the line (what I think was intended by the example), you'd do this:
the ! here means access external utility
 
 
----
 
To remove duplicate lines from a sorted list (make each line unique):
 
   
 
<pre>
 
<pre>
 
:sort /^.*|/
:%s/^\(.*\)\(\n\1\n\)\+/\1/g
 
 
</pre>
 
</pre>
 
works better in windows
 
<pre>
 
:%s/^\(.*\)\(\n\1\)\+/\1/g
 
:g/^$/d
 
</pre>
 
 
:Note that the ''todo'' at the top of these Comments points out we should all be using Vim 7 now, and that makes this tip obsolete.
 
:To sort lines and remove duplicates in Vim 7:<pre>:sort u</pre>
 
 
----
 

Revision as of 19:48, 2 January 2014

Tip 1166 Printable Monobook Previous Next

created 2006 · complexity basic · author Robert Stovall · version 7.0


Vim has a very powerful built-in sort utility, or it can interface with an external one. In order to keep only unique lines in Vim, you would:

:{range}sort u

Yes, it's that simple.

You could create a range in advance, such as 'a,. (from mark 'a' to the current line) or you could create one on-the-fly using visual selection by pressing ':' in visual mode, after selecting the text you wish to sort, to get a range of '<,'> on the command line.

If you like using an external sort utility instead, you can do it just as easily. For example, Unix sort, removing duplicate lines:

:{range}!sort -u

Many other systems also have an external sort utility, but the options and capabilities will differ. It is probably better to use the built-in Vim sort unless you are looking for a specific feature of the external sort (or using an old Vim without the :sort command).

Examples

Sort in reverse

:%sort!

Sort, removing duplicate lines

:%sort u

Sort using the external Unix sort utility, respecting month-name order

:%!sort -M

("respecting month-name order" means January < February < ... < December)

Numeric sort

:sort n

(this way, 100 doesn't precede 20 in the sort)

Sort subsections independently, in this example sort numbers between "start" and "end" markers

:g/start/+1,/end/-1 sort n

See also

References

Comments

 TO DO 

  • Probably need some general :sort command info.
  • Give examples of numeric sort and using regex sort.
  • Clean up my "see also" list. It's useful now for a comprehensive list of related tips, some of which need work. At least should add a note on point of tip.
  • If we're going to mention an external sort tool, we may as well include the following with a brief explanation. Vim could do this, but only with a complex regex. Or perhaps better, mention it in VimTip374 or VimTip923 in "see also". -k2 sorts on the second field (word by default).
:!sort -k2

This misguided snippet was added recently:

delimit the column using some char here I have | symbol as delimiter, once did with that you can use below command to sort specific column use -n if u want to sort numeric and its working on some version of vi and not on ubuntu vi :(
/|.*|/ | sort
used to match a patern |.*| used to match words delimited between || and | as piping commend and sort to sort

This is wrong and should never work. Here's what it is actually doing:

/|.*|/: jump to the next line that has two '|' characters in it, anywhere

|: command separator, this lets you start a new command on the current line

sort: do a default sort of the entire buffer

Basically this is the equivalent of typing :%sort.

Now, what you CAN do, is provide a pattern that the :sort command will skip over and ignore at the start of every line while sorting. For example, to sort based only on text after the last '|' character on the line (what I think was intended by the example), you'd do this:

:sort /^.*|/