Careful when using time. The bash built-in called time times 1 shell
statement (including pipes). The binary in /usr/bin/time only times the
command given - it does not span pipes.
[EMAIL PROTECTED] wrote:
> On Tue, Jul 29, 2008 at 02:15:54AM -0500, Robert Citek wrote:
>
>> Are you sure
to
consume approximately the same amount of ram.
Ken
--- On Tue, 7/29/08, Stephen Woodbridge <[EMAIL PROTECTED]> wrote:
From: Stephen Woodbridge <[EMAIL PROTECTED]>
Subject: Re: [sqlite] db vs shell
To: "General Discussion of SQLite Database" <sqlite-users@sqlite.or
Do things improve any if you increase the temporary cache size?
Compile with -DSQLITE_DEFAULT_TEMP_CACHE=100 or something?
How much memory does the [sort] process consume in the shell
version? What percentage of records are being trimmed by the
first [uniq] in the pipeline?
Dan.
On Jul 29,
I'm seeing a similar speed different with the 3X performance difference:
[EMAIL PROTECTED]:~/work$ true && ( set -x
> sqlite3 sample.db 'create table bar (foo text)'
> seq -w 1 200 | sed 's/^/id/' > list.txt
> sqlite3 sample.db '.imp "list.txt" "bar"'
> time -p sqlite3 sample.db 'select
On Tue, Jul 29, 2008 at 4:25 AM, <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 29, 2008 at 03:27:20AM -0500, Robert Citek wrote:
>> real 3.25
> ..
>> real 22.48
>
> I'm seeing the second being twice as -fast- as the first one here, still.
I don't follow. 22/3 ~ 7. Or do you mean when you run the
On Tue, Jul 29, 2008 at 03:27:20AM -0500, Robert Citek wrote:
> real 3.25
..
> real 22.48
I'm seeing the second being twice as -fast- as the first one here, still.
How many CPU cores are in your testing machine? Parallel execution
-might- explain the difference.
Cheers, Peter
2008/7/29 Robert Citek <[EMAIL PROTECTED]>:
> On Tue, Jul 29, 2008 at 2:35 AM, <[EMAIL PROTECTED]> wrote:
>> On Tue, Jul 29, 2008 at 02:29:53AM -0500, Robert Citek wrote:
>>> $ sqlite3 -version
>>> 3.4.2
>>
>> On 3.4.0 and 3.5.9 here, the pure-SQL version is -much- faster than the shell
>> pipe.
On Tue, Jul 29, 2008 at 2:35 AM, <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 29, 2008 at 02:29:53AM -0500, Robert Citek wrote:
>> $ sqlite3 -version
>> 3.4.2
>
> On 3.4.0 and 3.5.9 here, the pure-SQL version is -much- faster than the shell
> pipe. Could you tell us more about the contents of your
On Tue, Jul 29, 2008 at 02:29:53AM -0500, Robert Citek wrote:
> $ sqlite3 -version
> 3.4.2
On 3.4.0 and 3.5.9 here, the pure-SQL version is -much- faster than the shell
pipe. Could you tell us more about the contents of your database?
Cheers, Peter
___
On Tue, Jul 29, 2008 at 2:23 AM, <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 29, 2008 at 02:15:54AM -0500, Robert Citek wrote:
>> Are you sure time ignores everything after the pipe?
>
> Seems to depend on shell version - when I tested it here it definitely
> ignored everything after. Yours seems to
On Tue, Jul 29, 2008 at 02:15:54AM -0500, Robert Citek wrote:
> Are you sure time ignores everything after the pipe?
Seems to depend on shell version - when I tested it here it definitely
ignored everything after. Yours seems to do the right thing, which makes
your sqlite issue an interesting
On Tue, Jul 29, 2008 at 1:31 AM, <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 29, 2008 at 01:26:54AM -0500, Robert Citek wrote:
>> Why the difference in time?
>
> Your first test is only measuring how long sqlite needs to 'select foo from
> bar';
> all the commands after the pipe are ignored by
On Tue, Jul 29, 2008 at 01:26:54AM -0500, Robert Citek wrote:
> Why the difference in time?
Your first test is only measuring how long sqlite needs to 'select foo from
bar';
all the commands after the pipe are ignored by 'time'.
Try this: time -p sh -c "sqlite3 sample.db 'select foo from bar ;
Was doing some DB operations and felt they were going slower than they
should. So I did this quick test:
$ time -p sqlite3 sample.db 'select foo from bar ; ' | uniq | sort |
uniq | wc -l
209
real 5.64
user 5.36
sys 1.51
$ time -p sqlite3 sample.db 'select count(distinct foo) from bar ; '
14 matches
Mail list logo