CHANGELOG.md in alf-0.15.0 vs CHANGELOG.md in alf-0.16.0

- old
+ new

@@ -1,482 +1,19 @@ -# 0.13.0 / FIX ME +# 0.16.0 - 2014-02-18 -* Ruby 1.8.x is no longer supported. Please upgrade. +* Fixed SQL compilation when projecting over set operators (minus, union, + intersection) +* Fixed SQL compilation of join with DEE/DUM. +* Removed `is_table_dee` attribute when projecting all attributes away from + a SQL data source. +* Fixed compilation of shortcut operators involving complex expressions. +* Fixed Tuple() that now accepts no arg and returns an empty tuple. +* Fixed "comparison failure" errors in the presence of `nil`. You should NOT + use `nil`, but robustness is probably necessary anyway. +* Added the ability to load a .rash file through Path#load. Result is an array + of hashes. +* Fixed default signature of Hierarchize. +* Added a basic `image` operator as a shortcut expression. -## Enhancements of the shell interface +# 0.15.0 - 2013-11-01 -* When used in shell, the default database is set to the current folder instead of the embedded suppliers and parts example database. This saves you from having to use 'alf --db=.' everytime you want to use .csv or .rash files as base relations. An --examples option allows easily setting the embedded database as default one. - -* When used in shell, the default rendering format is set to --text if the standard output is detected to be a tty. This saves you from having to use 'alf ... | alf show' too many times. The behavior of alf in shell, 'alf show' in particular, might be broken for you (see below). Thanks go to @eregontp for this suggestion! - -* Added --json output format. - -## Enhancements of the ruby interface - -* Added Alf.connect for connecting to databases in the easiest possible way: - - Alf.connect("somewhere/to/a/folder") do |conn| ... end - Alf.connect("database.sqlite3") do |conn| ... end - Alf.connect("postgres://user:password@host/database") ... - Alf.connect(adapter: "sqlite", database: "test.sqlite3") ... - -* Evaluation of queries are now sent to a connection specifically: - - Alf.connect(...) do |conn| - conn.query{ (restrict :suppliers, ->{ status > 10 } )} - end - -* Alf::Relation now respond to aggregation functions with an object-oriented syntax: - - Relation(...).sum{ qty } - -* Alf::Relation(...) (aliased as Alf.relation) now recognize IO objects and Path.like instances and load them with available readers in the easiest possible way: - - Alf::Relation("path/to/a/suppliers.csv") - # => Alf::Relation[...] - -* Alf::Relation(...) ensures that attribute names are Symbols and symbolize them if needed. - -* Added Relation#tuple_extract, aliased as Relation#tuple! that returns the only tuple of the relation, or raises a NoSuchTupleError if no tuple or more than one. - -* Added Alf::Tuple(...) that behaves similarly to Alf::Relation(...) but for tuples. - -* Added Alf.reader as a convenient shortcut for Alf::Reader.reader. - - Alf.reader("path/to/a/suppliers.csv") - # => #<Alf::CSV::Reader:0x007fd554058440 ...> - -* Added Alf::Reader#path that always returns a Path instance, unless the reader operates on an IO/StringIO. Use Alf::Reader#input to get the source passed at construction. - -* All queries as well as tuple expressions (in restrictions, extensions, summarizations, etc.) are now evaluated in a cleaner and extended scope through a BasicObject extended with all database helpers. This has multiple advantages and one drawback: - * It allows you to have all database helpers available in those expressions. - * You no longer have to worry about name clashes with Kernel's methods. - * Kernel's functions are no longer accessible whithout prefixing with ::Kernel. - -* In sync with the previous point, Relation (the class), DUM and DEE are now defined globally (unless you define `ALF_NO_CORE_EXTENSIONS` before loading Alf). Those constants can thus be safely used in query expressions without experiencing a NameError. - -* Added a JSON renderer. - -## Bug fixes - -* The Aggregator class, Summarization type and Summarize operator have been made thread-safe through #happens that now takes a TupleScope instead of a tuple. - -* Sequel::Connection#relvar correctly raises a NoSuchRelvarError if no table can be found. - -## Broken stuff - -* The `heading` operator has been renamed `infer-heading` (Heading -> InferHeading accordingly). - -* `Alf.lispy` has been removed from the public API. Please use `connect` instead: - - Alf.lipsy(db_params).query{ ... } - - becomes: - - Alf.connect(db_params) do |conn| - conn.query{ ... } - end - -* The Environment concept as been removed and replaced by Connection. That also means that `environment` readers and writers here have been replaced according to cases. Also, the --env option has been renamed to --db in the shell command. This is a major incompatible change of Alf internals that might break existing code that extends Alf::Environment, Alf::Reader or any subclass. - -* Connection#dataset has been replaced to Connection#relvar and now serves relation variables instead of pure iterators. - -* You now have to explicitely use 'alf show --text > ...' or 'alf --text ... > ' if you don't want ruby hashes to be outputted to output files. This is a consequence of tty detection that ensures a better shell experience. - -* Kernel's functions are no longer accessible in tuple expressions that are executed within a BasicObject's scope. - -* The syntax (Relation :relvar_name) is no longer supported as it does not denote a relation literal at all. - -* `Iterator#to_rel` has been removed. Use `Iterator#to_relation` instead. - -* Renderers are no longer able to coerce their input from a Symbol. - -* Renderer.renderers and Renderer.each_renderer have been removed. Use Renderer.all and Renderer.each instead, respectively. - -* Reader.readers has been removed. Use Reader.all or Reader.each instead. - -* Aggregator.aggregators has been removed. Use Aggregator.all or Aggregator.each instead. - -* TupleExpression#call and TupleExpression#[] have been removed. - -* TuplePredicate has been replaced by Predicate. The latter is no longer a subclass of TupleExpression. - -## Bug fixes - -* The backports gem is no longer required on ruby =1.9 - -# 0.12.2 / 2012-06-12 - -* Bumped and weakened backports dependency to '~> 2.6' - -# 0.12.1 / 2012-03-13 - -## Enhancements - -* Add `Alf::Relation()`, with the same semantics as the main `Relation()` function - added in 0.12.0. - -## Bugfixes - -* Ensure that `Relation()` reads a reader/operator only once. -* The Sequel environment now correctly uses the jdbc driver when attempting to connect - to a sqlite database/file. - -# 0.12.0 / 2012-02-09 - -* Add a Relation() toplevel method that mimics Array(), Integer(), and so on. - That method uses Tools::ToRelation which is a set of Myrrha coercion rules. - The Relation() method helps building relation "literals" (say) for common - cases, such as the following: - - Relation(:name => "Alf") - # => (Relation (Tuple :name => "Alf")) - - Relation([{:name => "Alf"}, {:name => "Myrrha"}]) - # => (Relation (Tuple :name => "Alf"), (Tuple :name => "Myrrha")) - - Relation(:name => ["Alf", "Myrrha"]) - # => (Relation (Tuple :name => "Alf"), (Tuple :name => "Myrrha")) - -# 0.11.1 / 2012-01-25 - -## Bugfixes - -* Fix Aggregator.stddev and Lispy.stddev that were missing - -# 0.11.0 / 2012-01-25 - -## Broken APIs (private sections only) - -* All pipe() methods have been removed and replaced by arguments taken at - construction time. This affects the implementation of operators and may - require changes to contributed readers and renderers. -* Alf::Operator::Base -> Alf::Operator::InstanceMethods -* Alf::Reader::Base -> Alf::Reader::InstanceMethods -* Alf::Rendered::Base -> Alf::Rendered::InstanceMethods -* Alf::Environment::Base -> Alf::Environment::InstanceMethods -* Alf::Aggregator::Base -> Alf::Aggregator::InstanceMethods -* Alf::Buffer has been removed as well as the Sorted specialization. -* Alf::Environment::Explicit has been removed as well as associated tools. -* Ordering#order_of has been removed -* Ordering#order_by has been removed -* Alf::Tools.coerce now raise a Alf::CoercionError instead of a Myrrha::Error - in case of coercion failure - -## On the development side - -* Bumped ruby.noe to 1.7.0 -* Bumped rspec to 2.8.0 -* Bumped sequel to 3.30.0 - -# 0.10.1 / 2011-08-31 - -## Miscellaneous enhancements - -* Added Variance and Stddev aggregation operators; they are available under - Aggregator.variance{} and Aggregator.stddev{}, respectively -* Added a --pretty option to 'alf', whose semantics is delegated to the output - renderer. Same option is available on 'alf show'. -* 'alf show' now accepts an optional ordering argument. This avoids explicitely - including non-relational sort invocations in your pipe (sorting should be - seen as a displaying issue) -* Added an options hash to text renderer, :float_format among others -* Added a --ff option to 'alf show', for float format in text rendering - -## Bugfixes - -* Alf::Environment::(Folder/Explicit)#dataset now correctly raises a - NoSuchDatasetError when the dataset cannot be found, as required by the - specification. -* Alf::Reader.reader now correctly returns a Rash reader when invoked on a - StringIO - -# 0.10.0 / 2011-08-15 - -## New recognized data sources - -* Alf now provides an Environment implementation on top of a SQL database. - This means that SQL tables can now be used as data-sources. This feature - relies on the sequel gem ('gem install sequel' is required), that drives - recognized SQL servers. Then (don't forget that ALF_OPTS also exists): - - % alf --env=postgres://user:password@host/database show table - -* Alf now recognizes and allows manipulating .csv files as first-class data - sources. CSV output is also supported of course. Under ruby <= 1.9, the - fastercsv gem is required ('gem install fastercsv' is required). Then: - - % alf restrict suppliers.csv -- "city == 'Paris'" (input) - % alf show suppliers --csv (output) - -* Alf now recognizes and allows manipulating .log files as first-class data - sources. This feature relies on request-log-analyzer gem that provides the - parsers that Alf uses, and the log formats it recognizes - ('gem install request-log-analyzer' is required). See examples/logs. - -## New operators and enhancements - -* A GENERATOR operator is introduced. It allows generating a relation with one - auto-number attribute, up to a given size. - -* A COERCE operator is introduced. It provides a quick way to obtain type-safe - relations from type-unsafe sources like .csv files. For example: - - % alf coerce mydirtyfile.csv -- name String price Float at Time - - The coerce operator is of course available in Ruby as well: - - (coerce "mydirtyfile.csv", :name => String, :price => Float, :at => Time) - -* The DEFAULTS (non-relational) operator now accepts default values as tuple - expressions. When used in shell, provided default values are now evaluated - that way. This allows specifying default values as being computed on the - current tuple. - -* Aggregations in the Lispy DSL must not be prefixed by Agg:: anymore. - -## Miscellaneous enhancements - -* Added 'alf --input-reader' to specify $stdin format (csv, rash, etc.) -* Added 'alf -Ipath' that mimics ruby's -I (add path to $LOAD_PATH before run) -* Lispy#run supports command arguments to be passed as a string -* Lispy#run supports piped commands, with '|' as in shell - -## Hurting changes to Lispy DSL (and therefore to Relation) - -* The attribute-name syntax of aggregation operators has been removed. The Agg:: - prefix must not be specified anymore. - - Agg::sum(:qty) # !! error !! - Agg::sum{ qty } # !! error !! - sum{ qty } # simply, and only! - -* The group aggregation operator has been removed. It will probably be replaced - in a future version. In the meantime, the GROUP relational operator allows - obtaining similar results. - -* Lispy syntax of CLIP has changed (when used with --allbut option) - - (clip :suppliers, [:name, :city], true) (before) - (clip :suppliers, [:name, :city], :allbut => true) (after) - -* Lispy syntax of DEFAULTS has changed (when used with --strict option) - - (defaults :suppliers, {:country => 'Belgium'}, true) (before) - (defaults :suppliers, {:country => 'Belgium'}, :strict => true) (after) - -* Lispy syntax of GROUP has changed (when used with --allbut option) - - (group :supplies, [:sid], :supplying, true) (before) - (group :supplies, [:sid], :supplying, :allbut => true) (after) - -* Lispy syntax of PROJECT has changed (when used with --allbut option) - - (project :suppliers, [:name, :city], true) (before) - (project :suppliers, [:name, :city], :allbut => true) (after) - -* Lispy syntax of SUMMARIZE has changed (when used with --allbut option) - - (summarize :supplies, [:qty, :pid], {...}, true) (before) - (summarize :supplies, [:qty, :pid], {...}, :allbut => true) (after) - -## Hurting changes in shell - -* The attribute-name syntax of aggregation operators has been removed - - sum(:qty) # !! error !! - sum{ qty } # works - -* Shell syntax of GROUP has changed (option separator before introduced name) - - % alf --text group supplies -- pid qty supplying (before) - % alf --text group supplies -- pid qty -- supplying (after) - -* Shell syntax of WRAP has changed (option separator before introduced name) - - % alf --text wrap suppliers -- city status loc_and_status (before) - % alf --text wrap suppliers -- city status -- loc_and_status (after) - -* Shell syntax of QUOTA has changed (--by and --order become pure arguments) - - % alf quota supplies --by=sid --order=qty -- position count sum_qty "sum{ qty }" (before) - % alf quota supplies -- sid -- qty -- position count sum_qty "sum{ qty }" (after) - -* Shell syntax of RANK has changed (--order becomes a pure argument) - - % alf rank parts --order=weight,desc,pid,asc -- position (before) - % alf rank parts -- weight desc pid asc -- position (after) - -* Shell syntax of SUMMARIZE has changed (--by becomes a pure argument) - - % alf summarize supplies --by=sid -- total_qty "sum{ qty }" (before) - % alf summarize supplies -- sid -- total_qty "sum{ qty }" (after) - -## Bug fixes - -* [In shell] Options are now correctly parsed in presence of option separators. - That is, every argument after a '--' separator is considered a non-option - argument. - -# 0.9.3 / 2011-07-23 - -## New operators (available both in shell and in Lispy DSL) - -* Added MATCHING and NOT MATCHING operators. These operators are useful - shortcuts for the following expressions. - - (matching l, r) := (project (join l, r), [l's attributes]) - (not_matching l, r) := (minus l, (matching l, r)) - - For example: - - # Give suppliers who supply at least one part - (matching suppliers, supplies) - - # Give suppliers who don't supply any part - (not_matching suppliers, supplies) - -* Added RANK operator, which is useful for for computing quota queries as - illustrated below. See 'alf help rank' for details. - - # Give the three heaviest parts - (restrict (rank :parts, [[:weight, :desc]], :pos), lambda{ pos < 3 }) - -## Enhancements when using Alf in shell - -* added 'alf -r', that mimics 'ruby -r' (require library before run) - -* When alf is invoked in shell (and only in this case), ALF_OPTS is used as - global options to apply as if they were specified inline: - - % export ALF_OPTS="--env=. --yaml" - % alf show suppliers - - is the same as - - % alf --env=. --yaml show suppliers - -* 'alf --help' now distinguishes experimental operators (quota in particular) - from those coming from the (much more stable) **Tutorial D** specification. The - former should be used with care as their specification may change at any - time. - -## Enhancements when using Alf in Ruby - -* Alf.lispy now accepts any argument recognized by Environment.autodetect; it - obtains its working Environment that way. Among others: - - Alf.lispy(Alf::Environment.folder("path/to/an/existing/folder")) - - is the same as: - - Alf.lispy("path/to/an/existing/folder") - -* Added Relation::DUM and Relation::DEE constants (relations of empty heading - with no and one tuple, respectively). They are also available as DUM and DEE - in Lispy functional expressions. - -* Added a Heading abstraction, as a set of attribute (name, type) pairs. - -## Internal enhancements (extension points) - -* The Reader and Renderer classes accept a Hash of options as third - constructor argument. These options can be used by extension points. - -* The Environment class now provides a class-based registering mechanism 'ala' - Reader and Renderer. This allows auto-detecting the target environment when - --env=... is used in shell. See Environment.autodetect and - Environment#recognizes? for contributing to this extension point. - -* Internals now rely on Myrrha for code generation. This means that all - datatypes can now be safely used in relations and dumped to .rash files in - particular. - -## Bug fixes - -* Added Relation#allbut, forgotten in two previous releases -* Fixed (join xxx, DEE) and (join xxx, DUM) -* Fixed scoping bug when using attributes named :path, :expr or :block in - Lispy compiled expressions (coming from .alf files) -* Fixed 'alf --yaml show suppliers' that renderer a --text table instead of - a yaml output -* Fixed bugs when using Date and Time attributes with .rash files -* Fixed bugs when using Date and Time attributes in restrict expressions - compiled from the commandline -* Fixed a few bugs when using attribute names that are ruby keywords - (restrict & extend) - -# 0.9.2 / 2011.07.13 - -# Bug fixes - -* Fixed the "alf show" command (undefined method `chain') - -# 0.9.1 / 2011.07.13 - -## Enhancements (public APIs) - -* Added the in-memory Alf::Relation data structure and associated tooling. - This allows using Alf in a object-oriented usual way, in addition to the - functional DSL: - - Alf.lispy.evaluate { - (join (restrict :suppliers, lambda{ status > 10 }), :cities) - } - - is equivalent to - - suppliers, cities = [...], [...] - suppliers.restrict(lambda{ status > 10 }).join(cities) - - see README about how to obtain suppliers and cities relations in the first - place. - -* Summarize now accepts a --allbut option, to specify 'by' attributes from an - exclusion perspective - -* .alf files are now evaluated in such a way that backtraces are "traceability - friendly" - -## Non backward-compatible changes to public APIs - -* Lispy#with has been removed because not being stable enough. The clean way - of reusing sub-queries is as follows (non purely functional, so far) - - kept_suppliers = (restrict :suppliers, lambda{ status > 10 }) - with_countries = (join kept_suppliers, :cities) - supplying = (join with_countries, :supplies) - (summarize supplying, - [:country], - :which => Agg::group(:pid), - :total => Agg::sum{ qty }) - -* As a consequence, named data sources (Symbols, like :suppliers above) are - now resolved at compile time, which is less powerful, yet much simpler and - sound. - -* Nest and Unnest have been renamed to Wrap and Unwrap respectively. This is - to better conform to **Tutorial D**'s terminology. - -* Lispy#chain was kept public in 0.9.0 by error and has been entirely removed - from the DSL. - -## Enhancements (internals) - -* Reader.reader delegates to Reader.coerce when its first argument is not - a String. This allows calling Reader.reader(args.first || $stdin) in quickl - commands for example. - -* Operator, Operator::Relational and Operator::NonRelational have a .each - class method that yields operator classes - -## Bug fixes - -* Fixed a bug that led to an Nil error when using unary operators on $stdin -* Fixed a bug when summarizing or sorting on Symbol attributes with ruby 1.8 -* Fixed numerous crashes under rubinius - -# 0.9.0 / 2011.06.19 - -## Enhancements - -* Birthday! +New birthday.