[SPARK-17587][PYTHON][MLLIB] SparseVector __getitem__ should follow __getitem__ contract

## What changes were proposed in this pull request?

Replaces` ValueError` with `IndexError` when index passed to `ml` / `mllib` `SparseVector.__getitem__` is out of range. This ensures correct iteration behavior.

Replaces `ValueError` with `IndexError` for `DenseMatrix` and `SparkMatrix` in `ml` / `mllib`.

## How was this patch tested?

PySpark `ml` / `mllib` unit tests. Additional unit tests to prove that the problem has been resolved.

Author: zero323 <zero323@users.noreply.github.com>

Closes #15144 from zero323/SPARK-17587.

1 files changed, 5 insertions, 5 deletions

diff --git a/python/pyspark/ml/linalg/__init__.py b/python/pyspark/ml/linalg/__init__.py
index 05c0ac862f..a5df727fdb 100644
--- a/python/pyspark/ml/linalg/__init__.py
+++ b/python/pyspark/ml/linalg/__init__.py
@@ -713,7 +713,7 @@ class SparseVector(Vector):
                 "Indices must be of type integer, got type %s" % type(index))
 
         if index >= self.size or index < -self.size:
-            raise ValueError("Index %d out of bounds." % index)
+            raise IndexError("Index %d out of bounds." % index)
         if index < 0:
             index += self.size
 
@@ -960,10 +960,10 @@ class DenseMatrix(Matrix):
     def __getitem__(self, indices):
         i, j = indices
         if i < 0 or i >= self.numRows:
-            raise ValueError("Row index %d is out of range [0, %d)"
+            raise IndexError("Row index %d is out of range [0, %d)"
                              % (i, self.numRows))
         if j >= self.numCols or j < 0:
-            raise ValueError("Column index %d is out of range [0, %d)"
+            raise IndexError("Column index %d is out of range [0, %d)"
                              % (j, self.numCols))
 
         if self.isTransposed:
@@ -1090,10 +1090,10 @@ class SparseMatrix(Matrix):
     def __getitem__(self, indices):
         i, j = indices
         if i < 0 or i >= self.numRows:
-            raise ValueError("Row index %d is out of range [0, %d)"
+            raise IndexError("Row index %d is out of range [0, %d)"
                              % (i, self.numRows))
         if j < 0 or j >= self.numCols:
-            raise ValueError("Column index %d is out of range [0, %d)"
+            raise IndexError("Column index %d is out of range [0, %d)"
                              % (j, self.numCols))
 
         # If a CSR matrix is given, then the row index should be searched