Merge pull request #226 from Joshua-Dias-Barreto/describe

Added describe method for numeric dataframes

Author: jecisc

src/DataFrame-Tests/DataFrameTest.class.st

@@ -1552,6 +1552,26 @@ DataFrameTest >> testDataTypesWithNil [
 	self assert: df dataTypes equals: expected
 ]
 
+{ #category : #tests }
+DataFrameTest >> testDescribe [
+
+	| dataFrame expected |
+	dataFrame := DataFrame
+		             withRows: #( #( 1 1 ) #( 2 nil ) #( 3 1 ) )
+		             columnNames: #( 'A' 'B' ).
+
+	expected := DataFrame withRows:
+		            #( #( 3 2 1 1 1 2 3 3 SmallInteger )
+		               #( 2 1 0 1 1 1 1 1 SmallInteger ) ).
+
+	expected rowNames: #( 'A' 'B' ).
+	expected columnNames:
+		#( 'count' 'mean' 'std' 'min' '25%' '50%' '75%' 'max' 'dtype' ).
+	expected at: 1 at: 9 put: SmallInteger.
+	expected at: 2 at: 9 put: SmallInteger.
+	self assert: dataFrame describe equals: expected
+]
+
 { #category : #tests }
 DataFrameTest >> testDetect [
 	| actual expected |

src/DataFrame/DataFrame.class.st

@@ -941,6 +941,29 @@ DataFrame >> defaultHeadTailSize [
 	^ 5
 ]
 
+{ #category : #statistics }
+DataFrame >> describe [
+	"Answer another data frame with statistics describing the columns of this data frame"
+
+	| content |
+	content := self numericalColumns collect: [ :column |
+		           {
+			           column countNonNils.
+			           column average.
+			           column stdev.
+			           column min.
+			           column firstQuartile.
+			           column secondQuartile.
+			           column thirdQuartile.
+			           column max.
+			           column calculateDataType } ].
+
+	^ self class
+		  withRows: content
+		  rowNames: self numericalColumnNames
+		  columnNames: #( count mean std min '25%' '50%' '75%' max dtype )
+]
+
 { #category : #accessing }
 DataFrame >> dimensions [
 	"Returns the number of rows and number of columns in a DataFrame"