h5 document is the fifth-generation version of the hierarchical format, a file format for storing scientific data and library files, developed by the US Center for Supercomputing Applications Center file format for large-scale data storage and organization.
H5 simplify the configuration file into two major types of objects:
1 data set dataset, is multi-dimensional arrays of the same type of data
Group 2 group, is a container structure can contain other data sets and groups, if a file stored in different types of data sets, the data sets to the management group uses
Intuitive understanding, we can refer to the file system, different files stored in different directories:
Directory is hdf5 file group, describing the classified information DataSet dataset, the dataset will be more effective to manage and divide by group
File is hdf5 file dataset, specific data representation
FIG relationship data set and is the group:
h5 is a real file hierarchy, the file system data type of formula. In addition there is the metadata in the data set, i.e., metadata
For each dataset, the addition to the data itself, there are a lot of this data set attribute information in hdf5 while supporting the attribute information corresponding to the stored data set, the set of all the attribute information is called The metaData, h5 is the FIG. up a dataset file
Pandas when read in hdf file specified key value refers hdf file Group,
1 read_hdf(path_or_buf, key=None, mode='r', **kwargs)
When the file to be read to obtain h5, it contains only a Group, the key parameter may be omitted
2 to_hdf(path_or_buf, key, **kwargs)
When writing a file h5, the group must be specified write, i.e., key value can not be omitted