pykusto

Adding a new feature:

• Allow using "shuffle hints" in summarize queries

Fix breaking changes that were issued after the last release (due to the azure-kusto-data version upgrade)

Update dependencies requirement of the package azure-kusto-data to its latest version.

Added functions: make_bag_if, make_set_if, make_list_if.

Added functions: take_any, day_of_week
Add option for installing without dependencies which are not needed in PySpark
Enforce type checks during tests
Join: allow providing a two columns in one call to 'on'

Support PySpark

1. Small change to authentication flow: default auth method is still Azure CLI, but the fallback to device auth has been removed. Instead, if you want device auth, initiate PyKustoClient with auth_method=KustoConnectionStringBuilder.with_aad_device_authentication.
2. Added functions: rand, has_any.
3. Update azure-kusto-data dependency to 2.1.1
4. Add support for Python 3.9

• Add methods: extract and extract_all
• Support pandas version 1.2.3

Allow returning all columns in the functions: any, arg_min and arg_max

Make the following public:

1. Table and Database classes
2. All expression types

1. New feature: perform retries upon transient failures.
2. Bug fix: stop suppressing exceptions raised while fetching in background.

Bug fix

1. Modify fetching behavior to avoid confusing behavior if items are requested before they are done being fetched
2. Add methods 'not_in' and 'has_not'
3. Add 'mod' parameter to hash function
4. Add 'case_sensitive' parameter to 'is_in' method
5. Fix 'is_in' method for non-litral arrays: now uses 'set_has_element' instead of 'contains'
6. Add reference to base practices in documentation
7. Reduce code duplication in tests

Breaking changes

1. Import style changed to not specify submodules. E.g. from pykusto import PyKustoClient instead of from pykusto.client import PyKustoClient
2. Internal facing classes were moved under _src to avoid being imported.
3. has_cs function deleted, instead use has with True as the second argument.

Other changes

1. A convenient list of all externally-facing classes as the autocomplete of from pykusto import
2. Update pandas dependency max version to 1.1.0rc0
3. Autocomplete of column names now also works for union of tables
4. Added pack_dictionary function

1. Add functions: all_of, any_of.
2. Allow providing multiple predicates to where (#87)
3. Fix logic for testing containment in dynamic types (#96)
4. Prevent accidental use of logical operators (#97)

1. Add pretty_render method
2. Allow adding a table name to a floating query

1. Change default authentication method to Azure CLI, with fallback to AAD device
2. Fix bug where operator overloading did not work with a primitive left operand
3. Auto-quote column name if it contains only digits
4. Update dependencies versions