Yes, that could be even more effective.

On Tue, Jul 21, 2015 at 2:09 PM, Erik Schnetter <[email protected]> wrote:

> HDF5 file support compression. This is enabled via a flag when writing the
> file; when reading, it is automatically decompressed. I assume that
> compression would greatly reduce the file size.
>
> -erik
>
> On Tue, Jul 21, 2015 at 1:21 PM, Stefan Karpinski <[email protected]>
> wrote:
>
>> In your example data, each value is represented with two bytes: one for
>> the value, one for a comma or newline. Each Int64 value is 8 bytes. If all
>> your values are between 0 and 255, you could use UInt8 to represent them
>> and cut the size in half.
>>
>> On Tue, Jul 21, 2015 at 1:16 PM, paul analyst <[email protected]>
>> wrote:
>>
>>> I have data in txt file, some milons like this:
>>> 0,0,2,0,0,0,0,0,0,0,0,0,0,0,0,0,0
>>> 0,0,0,0,0,0,0,2,0,0,0,2,0,0,0,0,1
>>> 0,0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,1
>>>
>>> Coding win1250.
>>>
>>> size of dane.txt is 1.3 GB
>>>
>>> D=readcsv("dane.txt")
>>> k,l=size(D)
>>>
>>> using HDF5, JLD
>>> hfi=h5open("D.h5","w")
>>> close(hfi)
>>>
>>> fid = h5open("D.h5","r+")
>>> g = fid["/"]
>>> dset1 = d_create(g, "/D", datatype(Int64), dataspace(k,l))
>>> dset1[:,:]=D
>>> close(fid)
>>>
>>> After save to h5 file the file has 6.3 GB ? Why new file is 4 times
>>> biger?
>>> Paul
>>>
>>
>>
>
>
> --
> Erik Schnetter <[email protected]>
> http://www.perimeterinstitute.ca/personal/eschnetter/
>

Reply via email to