HDF5 MANIPULATOR

Simple manipulation on hdf5 files.

*Note: for files too big to fit in memory, use _big.py

Split

Split hdf5 file (requires the same no. of entries per dataset):

usage: ./split.py <options>

HDF5 MANIPULATOR (split)

optional arguments:
  -h, --help            show this help message and exit
  --prefix [path/to/filename_base]
                        prefix for splitted files (base on input file if not
                        defined)
  --filelist [path/to/filelist]
                        save output files list in txt file

required arguments:
  --input [path/to/input_file]
                        path to input hdf5 file
  --size [int]          number of entries per file

Example:

./split.py --input /path/to/my/data/data.hdf5 --size 100

will create /path/to/my/data/data_XXX.hdf5 files, each with 100 entries (the last one may have less no. of entries)

Merge

Merge hdf5 files (requires the same datasets, with the same shapes, in all input files):

usage: ./merge.py <options>

HDF5 MANIPULATOR (merge)

optional arguments:
  -h, --help            show this help message and exit

required arguments:
  --input [list of input files]
                        path to input hdf5 files to merge ('file1, file2,...'
                        will look for all files starts with file1 and file2
                        and ends with .hdf5)
  --output [path/to/filename]
                        path to output hdf5 file

Example:

./merge.py --input '/path1/basename1, /path2/basename2' --output merged.hdf5

will merge all files matching /path1/basename1* and /path2/basename2* into merged.hdf5 file

Extract

Extract chosen datasets from hdf5 file (requires the same no. of entries per dataset):

usage: ./extract.py <options>

HDF5 MANIPULATOR (extract)

optional arguments:
  -h, --help            show this help message and exit

required arguments:
  --input [path/to/filename]
                        path to input hdf5 file
  --output [path/to/filename]
                        path to output hdf5 file
  --keys ['key1, key2, ...']
                        list of datasets to be saved in the output file

Example:

./extract.py --input /path/to/input.hdf5 --output /path/to/output.hdf5 --keys 'dataset1, dataset2'

will extract dataset1 and dataset2 from input.hdf5 and save in output.hdf5

Combine

Save different datasets from different files into one output hdf5 (requires the same no. of entries per dataset within the file and one common key use for ordering):

usage: ./combine.py <options>

HDF5 MANIPULATOR (combine)

optional arguments:
  -h, --help            show this help message and exit
  --keys1 ['key1, key2, ...']
                        list of datasets to be extracted from the first input
                        file (use all if not defined)
  --keys2 ['key1, key2, ...']
                        list of datasets to be extracted from the second input
                        file (use all if not defined)

required arguments:
  --input1 [path/to/filename1]
                        path to first input hdf5 file
  --input2 [path/to/filename2]
                        path to second input hdf5 file
  --output [path/to/filename]
                        path to output hdf5 file
  --match [key]         the common key use to order data

Example 1:

./combine.py --input1 /path/to/file1 --input2 /path/to/file2 --output /path/to/output --match id

requires both input files have id key, and no other common keys; will create a file which contains all datasets from input files for all entries with matching ids
Example 2:

./combine.py --input1 /path/to/file1 --input2 /path/to/file2 --output /path/to/output --match id --keys1 'data1' --keys2 'data2, data3'

will create a file which contains data1 from file1, data2 and data3 from file2 (for all entries with matching idss)

Test: create_hdf5.py

Create several hdf5 files filled with random numbers, matrices etc.

Test: diff.py

Check if two hdf5 files are exactly the same.

Test: diff_big.py

Check if two hdf5 files are exactly the same. If single dataset is too big to fit into memory it can perform partial check [default] or full check.

Example 1:

./diff_big.py file1 file2

if some dataset is too big, it will check first 100 entries, last 100 entries, and random 100 entries.
Example 2:

./diff_big.py file1 file2 fullcheck

if some dataset is too big, it will check dataset entry by entry (takes a lot of time).

tomaszgolan / hdf5_manipulator Goto Github PK