v0.20.4

What's Changed

Significant performance improvements when downloading multiple files (particularly multiple small files)

PR - refactor: new parallel_download method using async by @imathews in #66

Full Changelog: v0.20.3...v0.20.4

v0.20.3

What's Changed

• fix: handle dates w/ SAS + Stata by @imathews in #65

Full Changelog: v0.20.2...v0.20.3

v0.20.1

What's Changed

• fix: error messaging in notebooks by @imathews in #64
• feat: add support for CSV files in notebook.create_output_table()

Full Changelog: v0.20.0...v0.20.1

v0.20.0

This release introduces a number of significant updates, including new ways to work with files, better integrations with SAS and Stata, and new methods for parallel data processing – alongside a series of significant performance improvements when streaming large tables and files

What's new

• table.file("path/to/file") -> you can now reference a file within a table directly by its name!
• table.to_directory() -> construct a Directory, which is a representation of the files within a table that follows standard filesystem tree semantics, and allows for powerful manipulation of files at scale.
  • Directory.download()
  • Directory.get()
  • Directory.list()
  • Directory.mount()
• Calling file.open() now behaves the same as opening a file on the filesystem. E.g., with file.open() as f: ...
• Redivis filesystems are automatically registered with fsspec. When using compatible libraries, you can reference Redivis files via the URI scheme: redivis://table_ref/path/to/file.
• New table.to_read_streams() method returns a list of readable data streams that can be used for parallel processing.
• New table.to_sas() and table.to_stata() methods streamline the ability to load data directly into SAS / Stata within compatible notebooks.
• New table.update_variables(), dataset.update_variables(), and workflow.update_variables() methods allow for the bulk management of variable metadata across a table, dataset, or workflow.
• All exceptions raised by Redivis now inherit from class redivis.exceptions.RedivisError, with additional context based on the type of exception (docs).

Deprecations

• file.stream() has been renamed to file.open(). Use of the stream method will print a deprecation warning, and will be fully removed in a future version.
• table.download_files() has been superseded by table.to_directory().download(). Use of the old method will print a deprecation warning, and will be removed in a future version.
• Previously, a file id was used in the file(<id>) constructor. This behavior has been deprecated, in favor of passing the file name / path directly. Using the file id will print a deprecation warning, and will be fully removed in a future version.

Breaking changes

• redivis.file() is not longer available. Please use `redivis.table("table_ref").file("path/to/file") instead
• table.list_rows(), which has been deprecated for a number of major versions, has now been removed and will raise an error. Use table.to_arrow_table().to_pylist() instead.

v0.18.18

This release introduces a new parameter interface for working with workflow parameters

• Parameter documentation
• Parameter interface
• New methods:
  • redivis.parameter()
  • workflow.list_parameters()
  • workflow.parameter()
  • parameter.create()
  • parameter.delete()
  • parameter.exists()
  • parameter.get()
  • parameter.get_values()
  • parameter.update

Full Changelog: v0.18.17...v0.18.18

v0.18.17

• refactor: improve type coercion when stream dataset tables

Full Changelog: v0.18.16...v0.18.17

v0.18.16

• perf: reduce update interval for progressbar (should particularly help in jupyter environments)

Full Changelog: v0.18.15...v0.18.16

v0.18.15

Minor bugfixes associated with API property assignment in methods from v0.18.14

Full Changelog: v0.18.14...v0.18.15

v0.18.14

What's changed

• transform.source_tables() is deprecated, please use transform.referenced_tables() instead to return a list of all tables referenced by the transform, or transform.source_table() to return the transform's source table.
• notebook.source_tables() is deprecated, please use notebook.referenced_tables() instead to return a list of all tables referenced by the notebook, or notebook.source_table() to return the notebook's source table.

What's new

A number of new methods facilitate programmatically managing your workflows

• transform.update(*, name, source_table)
• transform.referenced_tables()
• transform.source_table()
• notebook.update(*, name, source_table)
• notebook.referenced_tables()
• notebook.source_table()
• datasource.create()
• datasource.source_dataset()
• datasource.source_workflow()
• dataset.previous_version()
• dataset.next_version()

What's fixed

• variable.update() and table.update() now correctly return the table / variable instance (rather than None).

Full Changelog: v0.18.13...v0.18.14

v0.18.13

• fix(versions): fix bad NoneType reference in version.previous|next_version()

Full Changelog: v0.18.12...v0.18.13