Merge pull request #242 from Joshua-Dias-Barreto/untestedMethods

Added DataFrame>> #removeDuplicateRows

Author: jecisc

src/DataFrame-Tests/DataFrameTest.class.st

@@ -3694,6 +3694,26 @@ DataFrameTest >> testRemoveColumnsWithNilsAtRowOutOfRange [
 	self should: [ df removeColumnsWithNilsAtRow: 100 ] raise: SubscriptOutOfBounds
 ]
 
+{ #category : #tests }
+DataFrameTest >> testRemoveDuplicatedRows [
+
+|dataFrame|
+
+dataFrame := DataFrame withRows: #(
+		(Barcelona 1.609 true)
+   		(Dubai 2.789 true)
+   		(London 8.788 false)
+		(Barcelona 1.609 true)
+		(London 8.788 false)).
+
+	dataFrame rowNames: #(A B C D E).
+	dataFrame columnNames: #(City Population BeenThere).
+	
+dataFrame removeDuplicatedRows.
+self assert: df equals: dataFrame.
+	
+]
+
 { #category : #removing }
 DataFrameTest >> testRemoveRow [
 

src/DataFrame/DataFrame.class.st

@@ -1667,6 +1667,22 @@ DataFrame >> removeColumnsWithNilsAtRowNamed: rowName [
 	self removeColumnsOfRowElementsSatisfing: [ :ele | ele isNil ] onRowNamed: rowName
 ]
 
+{ #category : #removing }
+DataFrame >> removeDuplicatedRows [
+	"Removes duplicate rows of a dataframe except the first unique row"
+
+	| numberOfRows nextRowIndex currentRow row aSet |
+	aSet := Set new.
+	numberOfRows := self numberOfRows.
+	1 to: numberOfRows do: [ :currentRowIndex |
+		currentRow := self rowAt: currentRowIndex.
+		nextRowIndex := currentRowIndex + 1.
+		nextRowIndex to: numberOfRows do: [ :index |
+			row := self rowAt: index.
+			row values = currentRow values ifTrue: [ aSet add: index ] ] ].
+	^ self removeRowsAt: aSet
+]
+
 { #category : #removing }
 DataFrame >> removeRow: rowName [
 	"Removes the row named rowName from a data frame"