Added ability to remove nil from DataFrame and Series

[AtharvaKhare]

DataSeries can remove nils using the method removeNils
DataFrame can remove nils from rows and columns using methods:

• removeColumnsWithNilsAtRow:
• removeColumnsWithNilsAtRowNamed:
• removeRowsWithNilsAtColumn:
• removeRowsWithNilsAtColumnNamed:
  Added relevant tests for these methods

[AtharvaKhare]

Any idea why code coverage decreased? The code in question is already covered in a test

[olekscode]

This seems like a bad implementation:

DataFrameInternal >> removeRowsWithNilsAtColumn: columnNumber
    "Removes all rows having a nil value at the column columnNumber"

     | newContents rowsToDrop k |
    "rowsToDrop has 1 at i if i-th row needs to be dropped, else 0"
    rowsToDrop := (self columnAt: columnNumber) collect: [ :ele |
        (ele isNil) ifTrue: [ 1 ] ifFalse: [ 0 ] ].
    newContents := Array2D
        rows: (self numberOfRows - (rowsToDrop select: [ :ele | ele = 1 ]) size)
        columns: (self numberOfColumns).

     1 to: self numberOfColumns do: [ :j |
        k := 0.
        1 to: self numberOfRows do: [ :i |
            (rowsToDrop at: i) = 1 
                ifTrue: [ k := k + 1 ]
                ifFalse: [
                newContents at: i - k at: j put:
                    (contents at: i at: j) ]]].

     contents := newContents.

First you collect 0 and 1 (instead of true and false), then you check if element is equal to 1 (instead of having the boolean element). And so on

[olekscode]

I suggest that we create a method for removing rows that satisfy a certain condition (and the same for columns).
Then you call it with the condition that element at a certain column is nil

[AtharvaKhare]

@olekscode done! Any idea about coveralls? Tests are written :/