emacs-orgmode
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [O] searching for csv utilities


From: Nick Dokos
Subject: Re: [O] searching for csv utilities
Date: Wed, 03 Jun 2015 13:21:01 -0400
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.0.50 (gnu/linux)

Jude DaShiell <address@hidden> writes:

> This is a piece of a modified ecm that may show what's going on.
>
> cut here.
> |----------------------+---------+--------+--------+--------|
> | Averages:            |         |        |        |        |
> | Counts:              |         |        |        |        |
> | Maximums:            |         |        |        |        |
> | Medians:             |         |        |        |        |
> | Minimums:            |         |        |        |        |
> | Modes:               |         |        |        |        |
> | Standard Deviations: |         |        |        |        |
> | Sums:                | 108.69) | 70.45) | 66.62) | 92.93) |
> |----------------------+---------+--------+--------+--------|
> #+TBLFM: @>$2..@>$>=vmean(@I..@>>;%.2f)
>

The syntax is wrong - the format has to go *outside* the parens
of the function. Also the row spec is almost certainly wrong -
something like the following is needed (but see below for a more
exact specification of what the table looks like and what the
actual TBLFM line should be):

#+TBLFM: @>$2..@>$>=vmean(@address@hidden);%.2f

The @address@hidden restricts the range of rows to be considered to be between
the first and the second hline. IOW, the assumption is that you have
three areas in your table: a header line, followed by an hline, a number
of rows with data, followed by an hline and a number of rows containing
calculated statistics, followed by an (optional) hline - something like
this:

| Date                 |    Sys |   Dia |   Pul | Sugar |
|----------------------+--------+-------+-------+-------|
| [2014-04-27 Sun]     |    125 |    88 |    78 |    92 |
| [2014-04-28 Mon]     |    102 |    88 |    86 |    92 |
|----------------------+--------+-------+-------+-------|
| Averages:            | 113.50 | 88.00 | 82.00 | 92.00 |
| Counts:              |      2 |     2 |     2 |     2 |
| Maximums:            | 125.00 | 88.00 | 86.00 | 92.00 |
| Medians:             |        |       |       |       |
| Minimums:            |        |       |       |       |
| Modes:               |        |       |       |       |
| Standard Deviations: |        |       |       |       |
| Sums:                |        |       |       |       |
|----------------------+--------+-------+-------+-------|
#+TBLFM: @>>>>>>>>$2..@>>>>>>>>$> = vmean(@address@hidden);%.2f :: 
@>>>>>>>$2..@>>>>>>>$> = vcount(@address@hidden) :: @>>>>>>$2..@>>>>>>$> = 
vmax(@address@hidden);%.2f

The only challenges left are to refer to the correct row and to find the
appropriate calc function for the statistic at hand.

Unfortunately, org mode is not particularly flexible in referring to
cells (although it is a *lot* more flexible than it used to be). In
particular, referring to the "Averages:" row can not be done simply
(afaik): it's the first row after the second hline, so you'd think that
you can refer to it as @II+1 but hline-relative references are not
allowed on the LHS, so the best that I could do is refer to it as the
eightth column from the bottom: @>>>>>>>> and similarly for the others -
looks like chicken scratchings (or what amounts to the same thing, Perl
:-) ).

With that caveat, the above TBLFM line will calculate the first three of
the eight statistics you want - the rest can be done similarly with the
calc functions vmedian, vmin, vmode, vstdev and vsum; but you can see
it's getting unwieldy, particularly since the whole thing has to be on
one line (or has that changed? I'm not keeping track any longer...)

I believe that you are (partially or completely?) blind, which makes
this all the more difficult: counting how many "greater than" signs are
in each of those row specs is a pain for anybody, but I can't even
imagine how a blind person would deal with it.

> I haven't even attempted the rest of the math since I have no way to 
> predict where any of the results will land.
> On Tue, 2 
> Jun 2015, Jonathan Leech-Pepin wrote:
>
>> Date: Tue, 2 Jun 2015 08:04:20
>> From: Jonathan Leech-Pepin <address@hidden>
>> To: Jude DaShiell <address@hidden>
>> Cc: Org Mode Mailing List <address@hidden>
>> Subject: Re: [O] searching for csv utilities
>> 
>> Hello,
>>
>> On 2 June 2015 at 07:44, Jude DaShiell <address@hidden> wrote:
>>
>>> | Date             |               Sys | Dia | Pul | Sugar |
>>> |------------------+-------------------+-----+-----+-------|
>>> | [2014-04-27 Sun] |               125 |  88 |  78 |    92 |
>>> | [2014-04-28 Mon] |               102 |  88 |  86 |    92 |
>>> | Averages:        | =$2=vmean(@<..@>) |     |     |       |
>>> #+TBLFM: $2=$2=vmean(@<..@>)
>>>
>>
>> The formula in question is the culprit in this case (at least as stated
>> there).
>>
>> : $2=$2=vmean(@<..@>)
>>
>> Second column is equal to the second column which is equal to the mean of
>> all the values in the second column (including the header "Sys").
>>
>> If you change the table as follows:
>>
>> | Date             |   Sys | Dia | Pul | Sugar |
>> |------------------+-------+-----+-----+-------|
>> | [2014-04-27 Sun] |   125 |  88 |  78 |    92 |
>> | [2014-04-28 Mon] |   102 |  88 |  86 |    92 |
>> |------------------+-------+-----+-----+-------|
>> | Averages:        | 113.5 |  88 |  82 |    92 |
>> #+TBLFM: @>$2..@>$>=vmean(@address@hidden)
>>
>> All the values will properly compute.  If you want to avoid the second
>> HLINE above Averages: then change address@hidden to =@>>= (penultimate row)
>>
>> Regards,
>> Jon
>>
>>
>>> This is a cut down version of my full record set.  Sometimes when I key
>>> formulas in I get ?ERROR back for a result after keying in c-c+c-c once
>>> I've completed the formula and hit tab.  If I do c-u+c-c+c-c that sometimes
>>> generated ?ERROR.  Other times I key in a formula and the cursor gets
>>> locked and I have to hit c-g to exit #+TBLFM: mode; I don't know what's
>>> actually happening when that situation arises since other than suddenly
>>> finding the cursor locked I can neither tell what state I'm in or if a few
>>> more keystrokes are needed or if I've generated an error situation.
>>>
>>>  --
>>>
>>>
>>>
>>

HTH,
Nick




reply via email to

[Prev in Thread] Current Thread [Next in Thread]