DataFrame
DataFrame copied to clipboard
DataFrame in Pharo - tabular data structures for data analysis
At the moment, we have three methods in DataFrame's `sorting` protocol: ```Smalltalk DataFrame >> sortBy: aColumnName. DataFrame >> sortBy: aColumnName using: aBlock. DataFrame >> sortDescendingBy: aColumnName. ``` We need to...
``` Smalltalk a := DataSeries withValues: #(1 2 2 3 5). a makeCategorical. a isCategorical. "true" b := a select: [ :each | each = 2]. b isCategorical. "false" ```...
Currently DataFrame only expects a FileReference as parameter to create an instance: ```smalltalk DataFrame readFromCsv: '/dir1/dir2/myfile.csv' asFileReference ``` It could be a nice time saver to let the DataFrame actually...
The "Very simple Example" section of the readme has an example of how to create a simple DataFrame, but no simple example of manipulating/querying it. Adding such example would make...
It would be nice to have groups defined in the BaselineOfDataFrame, to avoid loading tests.
I think that DataFrame is now mature enough to have its own logo :) Check out the logos of: - PolyMath: https://github.com/PolyMathOrg/PolyMath - Cormas: https://github.com/cormas/cormas - Pandas: https://pandas.pydata.org/ - Multiple...
To inspect a DataFrame, we create a Spec table: ```st inspectionItems: aBuilder | table | table := aBuilder newTable. table addColumn: (SpIndexTableColumn new title: '#'; sortFunction: #yourself ascending; beNotExpandable; yourself)....
It is possible to do operations such as divide between 2 DataSeries. In the case of a DataSerie containing nils, we consider that the return of the operation should be...
Currently data types are saved in a collection of the DataFrame. IMO it would make more sense to save it in each data series.