Working with HDF5 Tables

HDF5 tables are one-dimensional, extensible HDF5 datasets whose elements are of an HDF5 compound datatype. PyHexad supports compound types whose fields (“columns”) are scalar, pre-defined HDF5 datatypes (integers, floating-point numbers, strings). There are currently four Excel functions in PyHexad to support standard oerpations on HDF5 tables.

The first function, h5readTable, can be used to read an entire HDF5 table into Excel or just a subset of columns or rows.

The second function, h5newTable, lets one create a new HDF5 table and customize its properties.

The third and fourth functions, h5appendRows and h5writeTable, allow one to append rows to an existing HDF5 table or overwrite existing rows.

Caution

Beyond the name table, HDF5 tables don’t share very much with their relational cousins. At the file level, they are stored as HDF5 datasets and that largely dictates their “semantics”, which means that reading from and writing to an HDF5 table is much more akin to accessing an array than querying or updating a relational table. See [PyTables] and [pandas] for better table abstractions on top of HDF5.

• Reading Tables: h5readTable
• Creating New Tables: h5newTable
• Appending Rows to a Table: h5appendRows
• Writing to a Table: h5writeTable