I want to thank you a lot because i made real progress to achieve my aim
and it's due to the pytables capabilities and the developper's community.
I got a problem to organise my data in the hdf5 file.
The problem is the data are linked together.
For exemple, i have a pressure depending from the time and space
I can easily store time in an hdf5 array.
What about space coordinates? Is it better to store it in a table with
each column corresponding to a dimension (x,y,z) ot to have independant
array (one for x values, one for y values, and one for z values)?
Furthermore, here is the complex aspect:
pressure is depending on 4 dimensions (time, x, y, z).
I want to store it in a table but i really don't know how because:
It is the combination of a a value for time a value for x a value for y
and a value for z that gives me a pressure value.
If i have 4 values for time, 4 values for x, 4 values for y for values
for z, it makes me 1 024 values to store with hierarchy dependancies.
Has someone an idea about this problem or perhaps a solution?
Which is the problem by defining a table like:
coords = FloatCol(shape=4)
pressure = FloatCol()
and assign values to the coords column (in the four dimensions) and
pressure? Please, note that if you are afraid about data redundancy and
getting a large dataset, you can activate compression (LZO if you need
high speed access to data, ZLIB or BZIP2 if you want to compress more).
You will be pleased seeing how small can be the resulting data file in
such a case.
Get latest updates about Open Source Projects, Conferences and News.