NEWS.md

SCDB (development version)

Improvements and Fixes

• delta_export and delta_load now accepts db_table_id of class Id without error.

• delta_load now always writes integer n_insertions and n_deactivations to the logs table (#221).

SCDB 0.6.0

Improvements and Fixes

• delta_export() and delta_import() functions are added (#213) These functions facilitate incremental backups and batch migration of a historical table created with update_snapshot(). See vignette("delta-loading") for usage details.

• Bug fixes in update_snapshot():

  • Fixed a bug records were being incorrectly deactivated (#214).

SCDB 0.5.2

Improvements and Fixes

• Bug fixes in update_snapshot():

  • Fixed a bug where the process would unexpectedly hang (#192).

  • filter is copied to both source and target connections if on different connections (#201).

• get_tables() no longer fails for some PostgreSQL configurations (#198).

• filter_keys() no longer fails if filter list has zero rows (#200).

• db_timestamp() now always casts to timestamp (#202).

SCDB 0.5.1

BREAKING CHANGES

• update_snapshot() no longer collapses continuous records by default (#183, @kjakobse). To preserve this behavior, use collapse_continuous_records = TRUE. In return, update_snapshot() is now faster by default.

Improvements and Fixes

• Intermediary tables are now being cleaned up (#174).

• update_snapshot() now handles the given timestamp more consistently (#187). Thanks to @kjakobse for discovering the issue.

• ?Logger$finalize() is now a private method (#180). Required as of R6 v2.4.0.

SCDB 0.5.0

New features

• Added function create_index() to allow easy creating of an index on a table (#137).

Improvements and Fixes

• update_snapshot() has been optimized and now runs faster on all the supported backends (#137).

• *_joins() can now take dplyr::join_by() as by argument when no na_by argument is given (#156).

• SCDB has been made backwards compatibility to R >= 3.6 (#164).

• nrow() now always returns integers (#163).

• interlace_sql has been fully deprecated and removed (#169).

Documentation

• A vignette including benchmarks of update_snapshot() across various backends is added (#138).

SCDB 0.4.1

Improvements and Fixes

• Logger now correctly writes to the "catalog" field on backends that support it (#149).

• get_schema() now correctly returns the temporary schema on PostgreSQL backends (#139).

• get_tables() now returns catalog on DuckDB backends (#145).

• unique_table_names() now uses random alphanumerics to form the unique name instead of tracking via options (#158).

Documentation

• Deprecated check_from argument no longer used in dbplyr calls (#136).

Testing

• Improved tests for get_tables() (#145).

SCDB 0.4.0

BREAKING CHANGES

• Table identification is now more specific (#93).

  Most SCDB functions allow for tables to be specified by a character representation of "[catalog].schema.table".

  Before, if no schema was implied in this context, SCDB would attempt to match the table among both permanent and temporary tables.

  Now, it will always assume that a lack of schema means the default schema should be used. This is also the case if DBI::Id() is used without a schema specification.

• The show_temporary argument of get_tables() is now a simple logical (#93).

  In addition, schema is always returned in the list of tables (no longer NA for default schema).

• Tables created with create_table() will now be temporary or permanent dependent on the default value of DBI::dbCreateTable() (#93).

  If you wish to overwrite this, use ... arguments which are passed to DBI::dbCreateTable().

• The %notin% operator has been removed from the package (#96).

• The db_table_id argument in create_table(), get_table(), table_exists() and id() is renamed to db_table (#115). Any object coercible by id() can now be passed to these functions.

• The order of arguments in create_logs_if_missing() has been swapped to match the rest of the package (#96). The conn argument is now before the log_table argument.

• The arguments of Logger has been updated (#98):

  • db_tablestring is replaced with the db_table argument.

    This argument takes any input coercible by id() instead of only allowing a character string.

  • ts is replaced with the timestamp argument to align with update_snapshot().

• The order of input arguments to Logger is changed (#98).

• The interlace_sql() function is deprecated in favor of the S3 generic interlace() (#113).

New features

• Added support for DuckDB (#121).

• The S3 method as.character.Id() is added which converts DBI::Id() to character (#93).

• A new id.data.frame() which converts data.frame to DBI::Id() (#108). Useful in combination with get_tables(conn, pattern).

• A new get_catalog() function is added to give more specific table identification (#99).

• A new clean up function, defer_db_cleanup(), is added (#89).

  By passing a tbl_sql object to this function, the corresponding table will be deleted once the parent function exits.

• A new function, unique_table_name(), to generate unique table names is added (#89). This function is heavily inspired by the unexported dbplyr:::unique_table_name().

• A logger is introduced LoggerNull (#98):

  • Logger facilitates logging to file/console and logging to database.

  • LoggerNull is "no-logging" logger that can be used to suppress all logging.

• Added a set of helper functions to prevent race conditions when writing to data bases (#104).

  See lock_table() and unlock_table().

Improvements and Fixes

• Improvements for create_table() (#93):

  • now writes the table if a remote connection is given. Before, it would only create the table with corresponding columns.

  • can now create temporary tables for Microsoft SQL Server.

• Improved checks on get_connection() (#83):

  • If given, host does not need to look like an IP address (e.g. "localhost" is not unrealistic).

  • A character input for port is allowed if it is a string of digits.

  • Now checks if timezone and timezone_out is an IANA time zone.

• digest_to_checksum() has improved performance on Microsoft SQL Server by use of the built-in HashBytes function (#97).

• table_exists() now correctly gives ambiguity warning on Microsoft SQL Server and PostgreSQL backends (#80).

• get_tables() now supports temporary tables for Microsoft SQL Server (#93).

• get_schema() has been updated (#107):

  • It will now always return a schema (either directly from the object or inferred by id()).

  • A temporary argument is added to get the temporary schemas from DBIConnections.

• id() now includes information of catalog in more cases (#99, #107).

• Fixed dplyr joins failing if testthat is not installed (#90).

• The footprint of update_snapshot() is reduced by cleaning up intermediate tables with defer_db_cleanup() (#89)

• update_snapshot() now attempts to get a lock on the table being updated before updating (#104).

• Logger$log_info() now uses message() instead of cat() to write to console (#98). The message written is now also returned invisibly.

Testing

• Added missing tests for create_logs_if_missing() (#93).

• Added missing tests for get_schema() (#99).

• Added missing tests for get_catalog() (#107).

• Improved tests for get_tables(), table_exists(), and create_table() (#93).

• Improved tests for Logger (#98).

SCDB 0.3

• Added support for Microsoft SQL Server using ODBC (#77).

Minor Improvements and Fixes

• Implementation of *_joins improved, now extending dplyr::*_joins rather than masking them (#77).

• Added S3 method for id.tbl_dbi(), returning a DBI::Id() instance matching the table (#72).

  • Calling id() on a tbl_dbi thus allows to retrieve a schema even when not initially given.

• Fixed update_snapshot() not working with a DBI::Id() instance as db_table argument (#72).

• Suppressed recurring messages from dbplyr >= 2.4.0 about table names containing . (#72).

• Added show_temp option to get_tables() to allow retrieving temporary tables (#72).

Other news

• Maintainer changed to Rasmus Skytte Randløv (@RasmusSkytte).

SCDB 0.2.1

Minor Improvements and Fixes

• SQLite connections now support schemata similar to other backends (#67).

• The package logo has been slightly altered to have a readable clock (#49).

• Added a vignette describing the concept of a slowly changing dimension using examples (#53).

• Added a Logger$finalize method, which removes the log_file in the database when not writing to a file (#66).

Other news

• Maintainer changed to Marcus Munch Grünewald (#59).

SCDB 0.2

Breaking changes

• update_snapshot() now take a Logger object through the logger argument instead of log_path and log_table_id arguments (#24).

• Logger\$log_filename has been changed to Logger\$log_basename to reduce ambiguity.

New features

• Package functions are now also tested with RPostgres::Postgres(), which is therefore now officially supported (#31).

• get_connection() shows a warning if an unsupported backend is used (#26).

• Increased flexibility for the Logger object (#21 #24):

  • A Logger instance may now be created with no arguments.

  • Suppress console output with output_to_console (TRUE by default).

  • If no log_path is set, Logger does not fail before trying to write to a file.

  • Logger\$log_realpath gives the full path to log file being written.

Minor improvements and fixes

• schema_exists correctly detects a schema with no tables (#30).

• db_timestamps now newer calls translate_sql with con = NULL (#37).

• Package description has been updated to not use a footnote on CRAN.

Known issues

• As schema_exists on an empty schema tests by creating a new table, this may cause issues if the user does not have sufficient privileges.

SCDB 0.1

Features

• Functions to handle database connections:

  • get_connection(), close_connection(), id()

• Functions to interface with database:

  • get_tables(), table_exists(), get_schema(), schema_exists()

• Functions to create "historical" tables and logs:

  • create_table(), create_logs_if_missing()

• Function to maintain "historical" tables:

  • update_snapshot()

• Functions to interface with "historical" tables:

  • get_table(), slice_time(), is.historical()

• Functions to facilitate faster joins with NAs on SQL backends:

  • full_join(), inner_join(), left_join(), right_join()

• Functions to manipulate tables on SQL backends:

  • filter_keys(), unite.tbl_dbi(), interlace_sql()

• A logging object to facilitate logging:

  • Logger()

• Function to generate checksums:

  • digest_to_checksum()

• Function to write timestamps to tables on SQL backends:

  • db_timestamp()

• Helper functions:

  • nrow() - database compliant nrow()

  • %notin% - negated %in%

Testing

• Most package functions are tested here.

Documentation

• The functions are fully documented.