Leonard,
Thanks for the great tips. But what's R?
On 10/02/2009 03:23 PM Leonard Mada wrote:
> Dear Ken,
>
> I came across a similar problem a few month ago.
>
> I always thought that the chances of hitting a rows limit are minimal,
> but then one day I got an error by Excel 2007 when I tried to
> import a csv-file.
>
> The csv file had slightly more than 2 million rows.
>
> Of course, the remaining spreadsheet applications failed
> miserably, too.
>
> The biggest problem was that even Excel - while loading
> 1 million rows, it still fails to graph that data.
> It will display only ~32,000 data-points, which was a
> very big disappointment.
>
> Fortunately, R solved very fast the problem.
>
> If you have a csv-file, load it into R:
> my.data <- read.csv("path/to/csv/csv_file.csv")
>
> Then plot the relevant variable:
> plot(my.data[[column_number]]
>
> where column_number is an integer equivalent to the column number.
>
> And by the way, R opened and processed the csv file orders
> of magnitude faster than equivalent spreadsheet programs.
>
> Sincerely,
>
> Leonard
>
>
> -------- Original-Nachricht --------
>> Datum: Fri, 02 Oct 2009 04:35:10 -0400
>> Von: ken <[email protected]>
>> An: [email protected]
>> Betreff: 2 Qs: size of tables & worksheets? graphed output to html?
>
>> I have two questions. Both pertain to a systems monitoring project I'm
>> developing for a college I work for. We want to monitor two linux
>> servers. Data (on mem, disk, cpu, etc. usage) will be collected every
>> ten seconds. This will then be graphed to show what's been going on
>> with the systems for various periods of time... like up to three weeks.
>> So this will be a lot of rows and columns. A piece of data every ten
>> seconds over three weeks comes close to 181,440 pieces of data. Let's
>> call this 200,000 rows. Multiply this by, say, ten data items each for
>> memory, swap, disk, and cpu and we have 8M data points, or cells. The
>> best info I got from google (thus far) was 2005 info which stated,
>> "GNUmeric could accept more than 256 columns and more of 65000 rows
>> (needs to be re-compiled)." Does anyone have better, more recent info
>> about gnumeric which would indicate that it could or could not handle
>> 200,000x40 cells? Better yet, what are gnumeric's data-holding
>> capacities?
>>
>> Secondly, users will query the database via a web form, which, once
>> submitted, produce a variety of graphs. Can gnumeric produce its graphs
>> as a separate image file suitable for insertion into a web page...?
>> like a jpeg or png file? And, if so, could those image files contain
>> borders noting, e.g., the days and hours along the x-axis and
>> percentages on the y-axis?
>>
>>
>> Of course the functionality I'm searching for would likely be more
>> readily found in a database app, like MySQL, but I've been using
>> gnumeric for a long time (though not its graphing capabilities), so am
>> more comfortable with it. If it provides the functionality I'm looking
>> for, then I can feel okay about bending the theoreticals.
>>
>>
>> TIA for your expert replies.
>>
>> --
>> War is a failure of the imagination.
>> --William Blake
>>
>> _______________________________________________
>> gnumeric-list mailing list
>> [email protected]
>> http://mail.gnome.org/mailman/listinfo/gnumeric-list
>
_______________________________________________
gnumeric-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/gnumeric-list
