Parquet to Hyper

Package to convert parquet files into a single hyper file.

Benchmarking

The package was benchmarked using a Latitude 3420 Dell Notebook with 16GB of Memory RAM, a 250GB SSD, and an i7-1165G7 CPU. The time may vary using different architectures. The table used for benchmarking contained 60 columns. Although the tests were carried with a maximum of 500 million of rows, the package supports higher amount of volume. The limitation is only to the size of a single parquet file (up to 30GB). For larger volumes, it's recommended to split them into multiple parquet files. Follow the results:

How to use

Installation

pip install parquet-to-hyper

Initializing object

from packages.hyper_file import HyperFile

parquet_folder = '/path/to/your/folder'             # The folder where the parquet files are
parquet_extension = 'parquet'                       # Optional. Don't use it if the parquet files has no extension
hf = HyperFile(parquet_folder, parquet_extension)

Create a single file

hyper_filename = 'path/to/your/db.hyper'            # Path to save hyper file with filename
rows = hf.create_hyper_file(hyper_file_name)
print(f'Hyper created with {rows} rows.')

Deleting rows from an existing hyper file

This function deletes rows based on a control column (date column) and the days to delete from current day.

hyper_filename = 'path/to/your/db.hyper'            # Path to load hyper file with filename
control_column = 'date_column'
days = 7
hf.delete_rows(hyper_filename)
print(f'{rows} rows were deleted.')

Appending rows from parquet into an existing hyper file

hyper_filename = 'path/to/your/db.hyper'            # Path to load hyper file with filename
rows  = hf.append_rows(hyper_filename)
print(f'{rows} were appended.')

Publishing hyper file into Tableau server

from packages.hyper_file import HyperFile

tsu = TableauServerUtils(tableau_address, token_name, token_value)
project_id = tsu.get_project_id(project_name)
tsu.publish_hyper(project_id, 'test.hyper')