All Versions
14
Latest Version
Avg Release Cycle
159 days
Latest Release
1129 days ago
Changelog History
Page 1
Changelog History
Page 1
-
v0.12.0 Changes
October 10, 2021β Added in 0.12.0
- β Add dataframe.GetGroups (@arjunmahishi)
- β Add Series.Slice (@jfussion)
- β Add csv lazy quote (@fredericlemoine)
π Changed in 0.12.0
- π series.Err is deprecated; use Error() instead
- π dataframe.Err is deprecated; use Error() instead
π Fixed in 0.12.0
- π Fix dataframe.GroupBy issue (@prliu)
- making series Order stable (@mcolosimo-p4)
-
v0.11.0 Changes
June 27, 2021β Added in 0.11.0
- Rolling window Mean and StdDev
- GroupBy and Aggregate
- Numeric column index
- Read HTML tables
- extra checks for TravisCI
- Combining filters with AND
- π User-defined filters
- Concatination of Dataframes
π Changed in 0.11.0
- π Make fixColnames faster
- π Use Go 1.16
- β‘οΈ Update dependencies
π Fixed in 0.11.0
- π Linter issues
- β Failing tests
-
v0.10.1 Changes
November 08, 2019π Fixed in 0.10.1
- π¨ LoadRecords printing type debug information
- Missing closing brackets in series.go
- π Fix gonum import path in dataframe_test
-
v0.10.0 Changes
November 08, 2019π Changed in 0.10.0
- π Merged dev branch changes from multiple collaborators (Sam Zaydel, Kyle Ellrott, Daniela Petruzalek, Christoph Laaber).
-
v0.9.0 Changes
October 03, 2017β Added in 0.9.0
- β Additional method to load arbitrary struct slices to DataFrames (Juan Γlvarez)
- π New LoadOption Names to set initial column names (Sander van Harmelen).
- π Parser option for csv delimiter (Kyle Ellrott)
- π New Describe method for reporting summary statistics (Daniela Petruzalek)
π Changed in 0.9.0
- π Improve the performance of multiple operations.
- π Code cleanup for better consistency (Sander van Harmelen)
- π Renamed 'Deselect' function to 'Drop' (Ben Marshall)
-
v0.8.0 Changes
December 12, 2016β Added in 0.8.0
- β Series.Order method and tests.
- β Series.IsNaN method and tests.
- β DataFrame.Arrange method and tests.
- β DataFrame.Capply method and tests.
- β DataFrame.Rapply method and tests.
- Benchmarks for several operations on both the
series
anddataframe
packages. - π Many optimizations that increase the performance dramatically.
- π New LoadOption where the elements to be parsed as NaN from string can be selected.
- Gota can now return an implementation of
gonum/mat64.Matrix
interface viaDataFrame.Matrix()
and load amat64.Matrix
viadataframe.LoadMatrix()
.
π Changed in 0.8.0
- elementInterface is now exported as Element.
- Split element.go into separate files for the implementations of the Element interface.
- π LoadOptions API has been renamed for better documentation via
godoc
. Series.Set
andDataFrame.Set
now modify the structure in place for performance considerations. If one wants to use the old behaviour, it is suggested to useDataFrame.Copy().Set(...)
instead ofDataFrame.Set(...)
.DataFrame.Dim
has been changed toDataFrame.Dims
for consistency with themat64.Matrix
interface.- π¨ When printing a large
DataFrame
now the behaviour of the stringer interface is much nicer, showing only the first 10 rows and limiting the number of characters that can be shown by line
β Removed in 0.8.0
- Some unused functions from the helpers.go file.
π Fix in 0.8.0
- π Linter errors.
- stringElement.Float now returns NaN instead of 0 when applicable.
- Autorenaming column names when
hasHeaders == false
now is consistent with the autorename used withdataframe.New
- π Bug where duplicated column names were not been assigned consecutive suffix numbers if the number of duplicates was greater than two.
-
v0.7.0 Changes
November 27, 2016β Added in 0.7.0
- β
Many more table tests for both
series
anddataframe
- Set method for
Series
andDataFrame
- When loading data from CSV, JSON, or Records, different
LoadOptions
can now be configured. This includes assigning a default type, manually specifying the column types and others. - π More documentation for previously undocumented functions.
π Changed in 0.7.0
- The project has been restructured on separated
dataframe
andseries
packages. - π
Reviewed entire
Series
codebase for better style and maintainability. DataFrame.Select
now accepts several types of indexes- Error messages are now more consistent.
- The standard way of checking for errors on both
series
anddataframe
is to check theErr
field on each structure. ReadCSV
/ReadJSON
andWriteCSV
/WriteJSON
now acceptio.Reader
andio.Writer
respectively.- β‘οΈ Updated README with the new changes.
β Removed in 0.7.0
- β Removed unnecessary abstraction layer on
Series.elements
- β
Many more table tests for both
-
v0.6.0 Changes
October 29, 2016β Added in 0.6.0
- InnerJoin, CrossJoin, RightJoin, LeftJoin, OuterJoin functions
π Changed in 0.6.0
- π¨ More code refactoring for easier maintenance and management
- β Add more documentation to the exported functions
- β Remove unnecessary methods and structures from the exported API
β Removed in 0.6.0
- colnames and coltypes from the DataFrame structure
-
v0.5.0 Changes
August 09, 2016β Added in 0.5.0
- Read and write DataFrames from CSV, JSON, []map[string]interface{}, [][]string.
- π New constructor for DataFrame accept Series and NamedSeries as arguments.
- Subset, Select, Rename, Mutate, Filter, RBind and CBind methods
- π Much Better error handling
π Changed in 0.5.0
- Almost complete rewrite of DataFrame code.
- π Now using Series as first class citizens and building blocks for DataFrames.
β Removed in 0.5.0
- π Merge/Join functions have been temporarily removed to be adapted to the new architecture.
- Cell interface for allowing custom types into the system.
-
v0.4.0 Changes
February 26, 2016β Added in 0.4.0
- Getter methods for nrows and ncols.
- π An InnerJoin function that performs an Inner Merge/Join of two DataFrames by the given keys.
- An RightJoin and LeftJoin functions that performs outer right/outer left joins of two DataFrames by the given keys.
- π A CrossJoin function that performs an Cross Merge/Join of two DataFrames.
- Cell interface now have to implement the NA() method that will return a empty cell for the given type.
- Cell interface now have to implement a Copy method.
π Changed in 0.4.0
- The
cell
interface is now exported:Cell
. - Cell method NA() is now IsNA().
- π The function parseColumn is now a method.
- A number of fields and methods are now expoted.
π Fixed in 0.4.0
- Now ensuring that generated subsets are in fact new copies entirely, not copying pointers to the same memory address.