A Tuesday 07 October 2008, Mico Filós escrigué: > Thanks Francesc for the information. > > I am a little puzzled, though. Isn't VLArray intended to represent > rows with a variable number of elements (ragged tables)? > In my case, what is variable is the number of rows (records), not the > number of columns; records have all fixed length.
That's right, but you asked to include sub-tables with a variable number of rows inside another table. This is what PyTables doesn't support. > In such case, is it possible to fill the entries of a table with > (sub)tables? [Fig. 2 of Dragan's article] It depends on the amount of data you have. If I understand well, Dragan's approach is to create a *different* Table object for every set of parameters. This is perfectly possible if you have a small set of different parameters (I mean, for all the *possible* combinations of parameters). However, if you have a lot of different sets of parameters (say, more than 10000), then you are going to need to create a lot of tables for keeping your observations, and this is not terribly efficent with PyTables/HDF5. In the case of having lots of possible combinations of parameters, another possibility would be to setup a single table with the next schema: A B C ntrial trajectory ------------------------------------------- a1 b1 c1 1 traj_array1 a1 b1 c1 2 traj_array2 a2 b1 c1 3 traj_array3 a2 b1 c1 4 traj_array4 i.e. the values in columns A, B and C would represent each possible combination of your parameters. This would let you to consolidate all your observations in a single table, and it is normally handier and more efficient to retrieve data from a single table than several. Moreover, you can minimize the overhead of the duplicated entries in A,B,C columns by setting the compression on. Other alternative would be to create a couple of tables, one for parameters, and the other for observations. The way to establish a relation one to many between both could be through a single additional VLArray dataset. So, for each row in the parameter table, there would be a row of values (with variable length) in the VLArray telling the rows of the observations that belongs to that parameter set. But, as you can see, the latter option is a bit more work than the first one, so my advice is to go with the former first and let PyTables/compression to do the dirty job for you. Hope this time things would be clearer now, -- Francesc Alted Freelance developer Tel +34-964-282-249 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users
