Changes:

  1. Refactor of tests to accommodate cran

Changes:

  1. Using readr column types to work around issues with inconsistent type conversion between DBI and JDBC drivers.

Bug fixes:

  1. Resolved issue where failed queries were being aborted inside the wrong connection in PooledConnectionHandler

  2. Refactored pooled connection handler to better ensure checkout connections are returned

Changes:

  1. More tidy cleanup of PooledConnectionHandlers to prevent leaked connections.

Changes:

  1. Changes made to make package CRAN compliant

  2. PooledConnectionHandler now uses withr::defer to automatically returned pooled connections returned with getConnection

Bug fixes:

  1. Fixed bug in uploads where empty tables are used to determine uploads (and can fail)

Bug fixes:

  1. Added type conversion checking on upload of data where columns are characters but interpreted as numeric from reading inserted data

Changes:

  1. For PooledConnectionHandler, added check to see if java stack size is set on unix systems before connecting stop overflow errors on rconnect platforms. Note that this solution will fail if RJava is called before the connection

Bug Fixes: 1. Removal of comment in DataMigrationManager sql that caused translation error for spark/databricks platforms

  1. The “optional” column in the model specification is now fully optional (when not present, all columns are assumed to be required)

Changes:

  1. Use of jdbc connection is default in pooled connection class

Changes:

  1. Tests for PooledConnection classes

Bug Fixes:

  1. Use of RPostgres updated.

  2. Use of dbplyr::in_schema for using dplyr::tbl in ConnectionHandlers

Changes:

  1. Allow PooledConnectionHandler classes to use DBI connections to bypass use of JDBC on systems where it may not be supported.

Bug Fixes:

  1. Fixed issue in platform specific migrations where SqlRender sometimes fails to add attribute when calling loadRenderTranslateSql

Bug fixes:

  1. Fixed issue with uploads results from removal of emptyIsNa property which is no longer used/required

Changes: 1. Added utility function grantTablePermissions to make it easier to grant select, delete, insert and update permissions for users on results database setups

  1. Added ResultExportManager class and utility to support standardized validation routines for exported data

  2. Allow packages to have an internal migration table prefix separate from the user defined one e.g my_study_sccs_migration

  3. emptyIsNa field is no longer required in specifications

  4. Purge all data supported in upload functionality (requiring user input)

Bug fixes:

  1. generateSqlSchema No longer requires primary key field to be lower case in results spec files, is now case-insensitive

  2. connectionHandlers now check sql string attribute to see if query needs translating, avoiding potential for errors caused by double translation

Changes:

  1. removed spuriously added createResultsDataModel function

  2. added QueryNamespace R6 class that allows users to define a table specification and then complete queries without always having to specify table names

  3. Added createQueryNamespace helper function to allow a variety of convenient ways to create query namespaces

  4. Added loadResultsDataModelSpecifications that loads results specs from a csv file and checks that the columns are correct

  5. Added some vignettes on usage of package features

Changes:

  1. Added support for validating results model schema data.frames

  2. Added function to generate sql for creating schemas from specification data frames

  3. Added unzip folder function

  4. Added robust upload function for directories of csv files conforming to model specifications

  5. Added vignette to describe creating results schemas and uploading data sets

Changes:

  1. Fixed support for newer versions of DatabaseConnector that have altered interface - this will mean this package now depends on DatabaseConnector version 6.0.0 and above. However, queryDb compatibility means that no code that uses these classes will need to be altered.

  2. Added support for ConnectionHandler to load tables with dbplyr/dplyr tbl interface (i.e. lazy loading). This means that dplyr style queries should be supported natively in Pooled and non-pooled connections.

Changes: 1. Added snakeCaseToCamelCase parameter to public in connectionHandlers so it can be defined once if required

  1. Added schema generator function that creates sql from csv files with table defs

Initial version