- Work on examples and doc.
- Clean up use of rank().
- Check for empty ... on more _se() methods.
- re-instate tests (not sure why they got left out).
- Move to tinytest.
- Add more related work links.
- More info in table handle.
- Isolate bquote a bit more.
- Don't import ApplyTo.
- Work on noSuggests issue.
- Better error message on named argument function.
- Don't optimize out select_columns as often.
- Add .(-x) notation to promote string values to names.
- Move wrapr to Depends.
- Work on query generation and clean up db adaptor path.
- Set minimum suggested rqdatatable version.
- Add many values vignette.
- Move to non-strict arrows.
- Make drop columns not strict.
- Better order documentation.
- Fix name substitution in immediate mode.
- Fix union_all.
- Allow ad-hoc (not-bquote) early LHS eval.
- Deal with bound variable types.
- Deal with degenerate partition in extend.
- Work on operator composition shortcuts.
- Tune yaml path.
- Fix project signatures.
- Fold consecutive select_columns nodes.
- Don't print quoted table name.
- Add arrow example.
- Add wrap/ex pattern.
- Remove relop_list paths.
- Fix drop_columns shortcut.
- Add YAML path for operator trees.
- Auto-register rqdatatable as default executor on load, if available.
- Fix column name quoting issue in operator presentation.
- Alternate operator names.
- example_employee_date() now accepts a wrapped database handle.
- move off legacy f_db signature.
- adjust license.
- Rename before join.
- Allow empty specifications in more situations.
- Remove str2lang() from vignette (method is not in older R)
- Improve expression re-mapping (include n()).
- Work on schema qualifications.
- Allow empty project.
- Better function re-mapping.
- Clean up test_set_indicator() error message.
- Don't override names() (messes up str()).
- Test checks that are correct when rqdatatable is attached.
- Add rquery substitution vingette.
- Remove non-ascii chars from source file to fix CRAN warning
- Fix dimnames().
- Add more argument checks.
- Fix temp-naming in joins.
- More tests.
- User controlled SQL-tree re-writing.
- More per-DB SQL rendering control.
- Move to tree-based parse representation (instead of flat).
- relop_list collector interface.
- Specialize materialze_node for lineage breaking.
- %%/MOD().
- Na row interfaces (un-publish tokenize_for_SQL()).
- Add mutable scratch area on nodes.
- versions without _nse() suffix.
- parse ^ power.
- reflexive arguments to non-sql nodes.
- Add .[]-variable deceleration notation.
- Narrow suggests.
- update docs.
- add order_expr.
- declare : as an inline operator.
- allow concatenating compatible pipelines.
- check for table name consistency.
- NULL/NA parsing.
- Force order limit to be >= 0.
- fix name lookup issue.
- add bquote() abstraction to extend_nse(), project_nse(), and select_rows_nse().
- fix column check on select_rows*
- confirm DBI on tests.
- Allow a bit more flexibility on ordering in extend.
- Start breaking up some recursive calls.
- Alternate data.table implementation path.
- lookup_by_column().
- Force parent.frame().
- Fix key_inspector_postgresql quoting issue.
- More tests.
- Export a default database description.
- Check more on project and extend interfaces.
- fix in-memory order_by!
- aggregate aliases.
- try to clean up immediate mode hooks a bit.
- Change to immediate execution.
- Substitute in values in presentation layer (rqdatatable depends on this).
- Fix name disambiguation.
- Better column dependency calculation.
- Re-map function names.
- Improve extend narrowing.
- Start on schema qualification.
- Get rid of %>>% and old dbi fn-names.
- Try to improve use of regexps.
- Add qlook().
- Add affine_transform().
- Documentation fixes.
- Make DBI suggested.
- Rename dbi_ prefixes to rq_ (for now have aliases from old to new).
- Per-connection options.
- Add expand_grid()/complete_design().
- Get rid of embedded data cases and print-execution.
- Move to new wrapr ops.
- Rationalize names of table description methods.
- Add rquery_executor hook.
- Don't default to setting up a temporary RSQLite database.
- Limit wrapr re-exports.
- Add ability to use %:=% in parsing.
- Move to executor options.
- Separate execute() and materialize() roles.
- Fix natural_join() semantics and make "by" a required argument.
- Fix if_else NA treatment.
- Fix orderby() limit bug.
- Better detection of column types: dbi_coltypes().
- Improve execute() performance in simple case.
- materialize_node().
- Minor documentation fixes.
- Add assign_slice(), if_else_op(), map_column_values(), and set_indicator().
- Bug fixes (esp on data.frame path).
- Work on printing/formatting.
- Improve column requirement checking.
- Improve column ordering.
- Make db_table() more prominent.
- Documentation improvements.
- Work around RPostgreSQL issue (dbTableExists does not work).
- More controls on dbi_ steps through package options.
- Push LIMIT deeper into pipelines.
- Add connection config tools.
- Fix re-run/re-create situations.
- Pipe database connection.
- Better quantile calc (assume window functions).
- Improve dependency decls.
- Further limit direct dependencies.
- Work with more DB drivers.
- SQL node for sets of columns.
- Move to wrapr 1.2.0
- Add wrapr_function.relop().
- Column checks on extend.
- Default sql_node to copying incoming columns.
- NULL column ops.
- rsummary().
- "Non SQL" nodes.
- New rquery_intro vignette.
- Improved help examples.
- Stricter argument checking.
- Be compatible with both wrapr 1.1.1 and 1.2.0.
- Fix "desc".
- Minor format change in rename (anticipate wrapr 1.2.0).
- Get ready for S3 wrapr_applicable and pipe_step (un-block wrapr 1.2.0 release).
- Don't store DB reference.
- Prepare for wrapr 1.1.0 features.
- Ad-hoc query modes.
- Fix statement partitioner.
- More flexibility on assignment notation.
- Initial experiment