DataCleaner vs Frictionless Framework

DataCleaner is an open-source data quality application for profiling, validating, and cleansing data. Its profiling engine identifies patterns, missing values, character sets, and other traits in data values. It works with CSV files, Excel spreadsheets, relational databases, and NoSQL databases. Users can create cleansing rules with search and replace, regular expressions, pattern matching, or custom transformations, and use internal or external reference data to check values. DataCleaner supports batch processing, visual workflows, standardization, validation, enrichment, scheduled runs, and desktop deployment on Linux, macOS, and Windows. Named integrations and connectivity options include Apache Hadoop, Apache Spark, Pentaho Data Integration, and Apache MetaModel. Developers can embed it in other applications or build plugins; the project also describes community extensions, integrations, and shared content. The code is licensed under LGPL. The community edition is free at free. The downloads page lists 5.9.0 as the latest release; release news says version 5.8.1 runs on Java 9 through 17 and is dated February 9, 2022.

Frictionless Framework is a free, open-source Python framework for describing, extracting, validating, and transforming tabular data. It can infer, edit, and save metadata such as table descriptions and field types, then validate tables, resources, and datasets and produce configurable reports. For data preparation, it can clean, reshape, and transfer data through pipeline and lower-level interfaces. It is available as a Python library and a command-line interface, and requires Python 3.8 or later. Supported formats include CSV, Excel, JSON, and SQL; CSV, Excel, and JSON are available by default. Its pluggable architecture supports custom checks and formats, as well as compressed files. The AWS plugin can read from and write to S3, while the CKAN portal feature can load and publish packages from a CKAN data management system. The Frictionless Framework plan costs nothing; installation is with pip. The software is distributed under the MIT License, whose terms permit use, copying, modification, distribution, sublicensing, and sale. The license provides no warranty.