I know the trick of the filter on the input data , but in reality the
problem is different :
1. For very large datasets ( 5 records ) with filtering of the dataset
we can only take small portions of data , making processing much faster .
Parthian you can operate almost in real time .
2. Often , when you take the data from the instrument, you should delete
records as containing incorrect information or taken in a time when the
experiment was not active. In this case to dissect the data is the only
way allow to obtain consistent diagrams .
3. I often work with diagrams that such are made with Temperature vs.
Electrical resistance , Pressure etc. In this case an experimental cycle
generates a graph in which are present, for the same temperature, a lot of
data . If you want to get a diagram, for example only with the data relating
to the period of descent of the temperature , in order to avoid side effects ,
I have to modify the file of input data . The thing is uncomfortable and then
totally impossible if that diagram I am using it to follow a real-time
experiment .
I understand that it may seem as strange required, but it is a typical request
made by those who work, as I, in the laboratory with experiments with very
long and variable mix and uses Veusz also as system analysis of the
experiment in replacing Labview and other similar programs .
Best regards
Adriano
On Thursday 05 February 2015 14:00:08 Steve Richardson wrote:
Ok, I see now what you mean, and how that could be really useful.
Have you tried using filtered datasets, in combination with the max
function? Specifically, I mean create a dataset using Data Operations
Filter Expression..., and then define a filter expression like:
(time (max(time) - 2))
to get the last 2 seconds of data. You'd have to filter both the time and
each data expression you want, but this might work.
I've attached an example.
Steve
On Thu, Feb 5, 2015 at 10:59 AM, Adriano Bassignana adri...@bassignana.eu
wrote:
I know it's possible to do so, but in a very uncomfortable and
impractical. As
you can see in the attached file, by now many times Veusz use as a monitor
of
the lab, and I must say that makes this work so well! The order to have
the
ability to filter, to the top, the dataset, allows to expand very much the
possibility of the program. For example, in this example that I enclose,
instead of seeing everything, I could just say, take all the records that
have
a timestamp no farther than 1000 seconds from now, at this point the
program
becomes a monitor! But I do not know where to put their hands in the code,
usually working in Java and do not know Python well as other :(
Then, in the case that there are experiments lasting several days, it
often
happens that you want to display only a time window or a specific group of
data. Sure, you can filter, but you have to parameterize everything and
the
speed of the system suffers a lot (sometimes also work with 0.1-1 million
records ...). If we could do the filtering directly on the RAW data, I
have a
simple and effective solution, avoiding the need to make a pre-processor
data.
Best regards
Adriano
On Thursday 05 February 2015 08:17:46 you wrote:
Hi,
You can easily achieve this functionality using python logical indexing.
For
example, to plot y vs x only where z 10, you can use expressions such
as:
x[z10]
y[z10]
for X data and Y data properties.
If your dataset name contains spaces or other special characters, you
will
need to use backticks like this
`dataset #2`[z10]
Hope this helps.
Steve
On Feb 5, 2015, at 4:09 AM, Adriano Bassignana adri...@bassignana.eu
wrote:
For many of my views would be very useful to have a global function of
dataset that allows you to filter the records in the data set as a
function for example of the values of a single record.
For example, let 's vole in the dataset display only records that have
the
values of a certain field of a certain value . You should be able
to
put that function as a filter on top ( Similar to the Filter field
currently present) . This way you can represent only the records that
are
considered correct within a certain experiment.
The thing you would do even more interesting if you could store the
filter
, maybe selectable with a pulldown menu , so that it can retrieve it
when
needed and then get different views of the data.
I think this feature is not complex , as it works on the entire
dadaset ,
but very useful for those who use , like me , this program to display
the
status of the experiments .
Best regards,
--
Adriano Bassignana
Advisory consulent on energy technologies
Via San Francesco 16
24061 Albano S.