-
Notifications
You must be signed in to change notification settings - Fork 75
SimpleFlatMapper ReleaseNotes
Arnaud Roger edited this page Sep 18, 2016
·
48 revisions
SimpleFlatMapper provides fast and easy to use mappers
- Big restruture to prepare for java9
- Csv and jdbc split in sfm-csv and sfm-jdbc module
- converter use ServiceLoader for extension
- package rename org.sfm to org.simpleflatmapper, org.sfm.utils to org.simpleflatmapper.util
- rename RowHandler to CheckedConsumer
- ...
- Fix issue with UUID key in Crud
- Jdbc support native uuid on postgres
- csv date parser supports multi format per column
- support datastax 3.1
- fix perf regression in csv mapper
- csv parser better buffer resize condition
- setup jdk 9 travis build
- fix .. when generating class name for asm mapper on Array
- Csv
- add ParallelReader
- Perf improvement on csv with quotes, slight decrease with no quotes
- CellConsumer is a Functionnal Interface
- Jdbc Crud
- better error message when primary key not defined
- support manual key definition
- support java date time for jooq integration
- Csv parsing perf improvement accross all strategy quotes/no quotes.
- Default value supported on all mapper.
- Csv parsing perf improvement.
- Param Name deductor printStackTrace
- Jooq Enum null Handling
CsvParser
- default value with no mapping does no need Date Format
- can specify a supplier in defaultvalueproperty
CsvParser
- parse directly from string
- trim spaces option
- DefaultValueProperty support
- mysql batch upsert/insert downsizing fix
- fix some error reporting for csv
- support value object deflation for crud
- better handling of value object reading and writing
- joda/java time support for datastax PS setter
- perf improvement for dynamic mapppers
- support Builder pattern for instantiation
- allow to specify instatiator Method or Builder Constractor or factory method
- support UUID for jdbc mapper ResultSet/PreparedStatement
- support UUID for csv mapper
- support Custom setter for PreparedStatement setter
- remove need for joda time runtime dep for PreparedStatement injection
- support java8 -parameters
- static method with generic type fix
- datastax crud support @table
- datastax mapping support @column
- jdbc crud support @table
- table name lookup
- add ttl and time out to datastax crud
- make it easier to customise CellWriter in CsvWriter DSL
- crud upsert support for mysql
- crud upsert support for postgresql
- crud batch insert/upsert uses insert multiple rows
- csv parser bug when unescaping a "" return " instead empty string fixed
- crud support mysql specific batch insert multiple values
- crud fix issue when generating queries for composite keys
- crud support batch create, update, read, delete using jdbc batching
- crud fix npe issue with generated keys but no consumer
- crud support auto increment column
- try to match param of the constructor with a getter when asm not present
- add crud support
- Read csv from file
- Support for named query from name parameter
- heuristic to extrapolate name from query when using ? placeholder
- Allow sfm bundle to see asm to be able to figure out constructor param name
- Fix mapper injector list index issue
- Fix asm linkage error on object with synthetic method
- Add support for array of SqlParameterSource for batch update
- add support for custom def on SqlParameterSource
- add support for spring SqlParameterSource
- fix bug in jooq record mapper provider
- static constant value for mapper to flat structure
- better object metadata building
- map object to PreparedStatement
- map object to Datastax PreparedStatement
- datastax UUID <-> String
- CsvParser String memory tripwire
- Datastax
- User Defined Type support
- List/Set/Mapper support
- Tuple support
- extract non java se support to sub modules
- add cassandra support, from benchmark there is about 20% speed increase compare to datastax mapper.
- fix bug on csv non asm error handling
- fix bug on csv parsing date in subobject
- fix bug on csv writing subobject
- better handling of 1 param object type
- csv escaping of line more consistent when customising end of line
- perf improvement
- java time and joda time support
- better configuration support
- append method now return this
- add csv writer
CsvWriter<DbObject> csvWriter = CsvWriter.from(DbObject.class).to(writer);
csvWriter.append(obj1);- fix concurrency bug on join and discriminator mapper.
- excel POI support see javadoc
- better handling of generic types
- big refactor to make it easier to create new mapper factory, builder on pull mapper
- support asm generation on all field mapper builder
- add Map, ?> mapping support
- better handling of generic parameter
- handle custom list/map impl
- add optional support
- add java 8 time support
- add factory method support
- better customisation of timezones and date formatter for joda and java8 time
- add asm class cache for jdbc mapper to avoid generating duplicated classes
- add asm class cache for csv handler to avoid generating duplicated classes
- reenable the asm handler generation with CsvParser DSL
- disable cell handler asm for parser dsl
- implement method sharding for better perd with big property number
- fix issue on csv mapper asm handler on object with more than 256 fields
- add threshold to switch to no asm over when faster
- better csv mapper performance even for non asm
- use asm csv handler see #116 benchmark results*
- beef up javadoc
- use speculative match on tuple and list, if no index are present will try map the prefix to the first available index
p_id, p_name, s_id, s_name
for a Tuple2 p will be associated to the first element and s to the second.
- does not wrap non runtime exception anymore use erasure trick to pass them through
- change all the package to optional in maven config, and made asm non optional
- fix issue with constructor injection and no asm
- check for null getter only if keys are defined
- travis build with fattuple revealed potential issue with internal janino field.
- public field asm support
- possibility to define keys on CsvParser DSL
List<Professor> list =
CsvParser.mapTo(Professor.class)
.addKeys("id", "students_id")
.forEach(reader, new ListHandler<Professor>()).getList()- 1-n join on Csv Mapper
- jOOL tuple support
- fasttuple support
- fix key predicate composition definition
- 1-n join with tuple as root
JdbcMapper<Tuple2<Professor, List<Student>>> mapper =
JdbcMapperFactory
.newInstance()
.addKeys("0_id", "1_id") // specify the keys to aggregate on and detect null
.newMapper(new TypeReference<Tuple2<Professor, List<Student>>>() {});- allow predicate on key to decide on the level it applies to.
JdbcMapper<Person> mapper = JdbcMapperFactory.newInstance()
.addColumnDefinition(
(key) -> key.getIndex() <= 2,
FieldMapperColumnDefinition.<JdbcColumnKey, ResultSet>key(
(pm) -> pm.getPath().startsWith("key.")
))
.newMapper(Person.class);JdbcMapper<Professor> mapper =
JdbcMapperFactory
.newInstance()
.addKeys("id", "students_id") // specify the keys to aggregate on and detect null
.newMapper(Professor.class);- inheritance support
JdbcMapper<JoinJdbcMapperTest.Person> mapper =
JdbcMapperFactoryHelper.asm()
.addKeys("id", "students_id")
.<JoinJdbcMapperTest.Person>newDiscriminator("person_type")
.when("student", JoinJdbcMapperTest.StudentGS.class)
.addMapping("person_type")
.addMapping("id")
.addMapping("name")
.when("professor", JoinJdbcMapperTest.ProfessorGS.class)
.mapper();- some bug fixes
- fix bug on csv parser with last cell if contains only one char
- better support for direct read type as root - String, Enum etc ...-
- 1-N support on jdbc mapper for setter and field injection path
JdbcMapper<ProfessorField> mapper = JdbcMapperFactory.newInstance()
.newBuilder(ProfessorField.class)
.addMapping("id")
.addMapping("name")
.addMapping("students_id")
.addMapping("students_name")
.joinOn("id");- support for easier generic type resolving with TypeReference support
- better toString output on columnDefinition, column keys and mapper
- add tuples for 2 to 32
- better signature for row handler
- Better generic support for TypeVariable
- add support for ignore
- add forEach on CsvParser DSL
- Change groupId to org.simpleflatmapper
JdbcMapper
- check null only id result return 0
- Calendar support
- support custom getter factory on column definition
- easier column definition composition
FieldMapperColumnDefinition.<JdbcColumnKey,ResultSet>identity()
.addRename("blop")
.addGetter(getter)
.addFieldMapper(fieldMapper)
.addGetterFactory(getterFactory)CsvMapper
- Calendar support
- Timezone setting support
- support CellValueReaderFactory on column definition
- easier column definition composition
CsvColumnDefinition.IDENTITY
.addDateFormat("yyyyMM")
.addRename("new_name")
.addCustomReader(reader)
.addCustomCellValueReaderFactory(cellValueReaderFactory)
.addTimeZone(tz);- can add column definition on Parser DSL
// dynamic mapper
CsvParser.mapTo(String.class, String.class)
.columnDefinition("1", CsvColumnDefinition.customReaderDefinition(cellReader));
// static mapper
CsvParser.mapTo(String.class, String.class)
.addMapping("0")
.addMapping("1", CsvColumnDefinition.customReaderDefinition(cellReader))- Joda DateTime, LocalDate, LocalTime, LocalDateTime supports
- can provide on CellValueReaderFactory
- can specify CsvColumnDefinition via a predicate
CsvMapperFactory
.newInstance()
.addColumnDefinition(
(k) -> k.getName().endsWith("_date"),
CsvColumnDefinition.dateFormatDefinition("yyyyMMdd")
).newMapper(MyObject.class)- add convenience method to override the headers
CsvParser
.mapTo(String.class, String.class)
.overrideWithDefaultHeaders()
.iterate(new StringReader("key,value\nvalue1,value2"));
CsvParser.mapTo(String.class, String.class)
.defaultHeaders()
.iterate(new StringReader("value1,value2"));- Joda DateTime, LocalDate, LocalTime, LocalDateTime supports
- can specify CsvColumnDefinition via a predicate
JdbcMapperFactory
.newInstance()
.addColumnDefinition(
(k) -> k.getName().endsWith("_date"),
columnDefinition)
).newMapper(MyObject.class)- allow for field customisation date format or custom reader
CsvMapperFactory
.newInstance()
.addColumnDefinition("date", CsvColumnDefinition.dateFormatDefinition("yyyyMMdd")
.newMapper(MyClass.class);
CsvMapperFactory.newInstance()
.newBuilder(MyClass.class)
.addMapping("date", CsvColumnDefinition.dateFormatDefinition("yyyyMMdd")
.mapper();- support tuple on the csv dsl
CsvParser.mapTo(String.class, Date.class).stream(reader);- allow for getter customisation
JdbcMapperFactory.newInstance().addCustomGetter("id", getter).newMapper(MyClass.class);- support java lang type as root object
CsvParser.mapTo(String.class).stream(reader);Query query = sql2o.open().createQuery("select * from table");
query.setResultSetHandlerFactoryBuilder(new SfmResultSetHandlerFactoryBuilder());
List<DbObject> dbObjects = query.executeAndFetch(DbObject.class);- replace non fluent call with fluent dsl see CsvParser How-to-parse-a-csv-file-in-java
- detect class loader visibility issues automatically
- fix issue with instantiator cache
- remove osgi services and activator as now class loader issue automatically detected
- disable asm in osgi as classloader issue of bundles being refresh it's not worth it for now
- allow change of row handler error handler
- tuple support
- support generic type resolution
- index discovery
- customizable separator and quote character for csv parser
- unescaping is now responsibility of the parser
- easier api to use the csv parser directly
- more stable performance accross csv size and version
- move buffer index to charconsumer in csv parser
- add streams and iterator to csvparser
- provide own impl of spliterator for faster streams
- refactor skip and limit for csv for faster parsing and easier maintenance
- add streams and iterator to jdbc and csv
- add JPA @Column support
- support for array of object
- support mapping sql Array to List
- add Jooq Integration
- add support for all resultset return types.
- restructure the packages to make javadoc more readable
- move mapper readme content to their own package
- JdbcMapper if a column is map to a property of object, it will try to look for one argument constructor that matches the column type.
- CsvMapper if a column is map to a property of object, it will use a one argument constructor if it's the only constructor.
- CsvMapper remove performance degradation with TieredCompilation
CsvMapper
- improve performance of unescaping.
- support CR CRLF and LF for end of line.
- handle space in header name.
- fix miss mapping issue when a mapping error occured.
- change default parser buffer size
- Asm
- fix concurrency issue that triggered LinkageError
- JdbcMapper
- increase perf to outperform Roma with the static mapper
- change benchmark to match roma result set loop
- fix for custom field mapping and primitive
- CsvMapper
- add dynamic csv mapper that configure it self based on the row
- add support for aliases and custom readers
- ability to customized date format
- add a Parsing Context to store the DateFormat per parsing.
- remove byte array based parsing as benchmark so it would only be rarely lead to perf gain
- fix for custom mapper
- simplify parsing code
- CsvMapper
- non boxing mapping
- constructor injection
- inner object
- direct byte parsing
- extract common part of MapperBuilder
- add Query DSL Jdbc support
- move test to orm-benchmark