Remove lazy parameters for GetRowCount (#1621)

* Remove lazy parameters for GetRowCount

* Address comments

Author: wschin

src/Microsoft.ML.Api/DataViewConstructionUtils.cs

@@ -397,7 +397,7 @@ protected DataViewBase(IHostEnvironment env, string name, InternalSchemaDefiniti
                 }
             }
 
-            public abstract long? GetRowCount(bool lazy = true);
+            public abstract long? GetRowCount();
 
             public abstract IRowCursor GetRowCursor(Func<int, bool> predicate, IRandom rand = null);
 
@@ -555,7 +555,7 @@ public override bool CanShuffle
                 get { return true; }
             }
 
-            public override long? GetRowCount(bool lazy = true)
+            public override long? GetRowCount()
             {
                 return _data.Count;
             }
@@ -654,7 +654,7 @@ public override bool CanShuffle
                 get { return false; }
             }
 
-            public override long? GetRowCount(bool lazy = true)
+            public override long? GetRowCount()
             {
                 return (_data as ICollection<TRow>)?.Count;
             }
@@ -735,7 +735,7 @@ public override bool CanShuffle
                 get { return false; }
             }
 
-            public override long? GetRowCount(bool lazy = true)
+            public override long? GetRowCount()
             {
                 return null;
             }

src/Microsoft.ML.Api/StatefulFilterTransform.cs

@@ -99,7 +99,7 @@ private StatefulFilterTransform(IHostEnvironment env, StatefulFilterTransform<TS
 
         public Schema Schema => _bindings.Schema;
 
-        public long? GetRowCount(bool lazy = true)
+        public long? GetRowCount()
         {
             // REVIEW: currently stateful map is implemented via filter, and this is sub-optimal.
             return null;

src/Microsoft.ML.Core/Data/IDataView.cs

@@ -82,17 +82,15 @@ public interface IDataView : ISchematized
         bool CanShuffle { get; }
 
         /// <summary>
-        /// Returns the number of rows if known. Null means unknown. If lazy is true, then
-        /// this is permitted to return null when it might return a non-null value on a subsequent
-        /// call. This indicates, that the transform does not YET know the number of rows, but
-        /// may in the future. If lazy is false, then this is permitted to do some work (no more
-        /// that it would normally do for cursoring) to determine the number of rows.
+        /// Returns the number of rows if known. Returning null means that the row count is unknown but
+        /// it might return a non-null value on a subsequent call. This indicates, that the transform does
+        /// not YET know the number of rows, but may in the future. Its implementation's computation
+        /// complexity should be O(1).
         ///
-        /// Most components will return the same answer whether lazy is true or false. Some, like
-        /// a cache, might return null until the cache is fully populated (when lazy is true). When
-        /// lazy is false, such a cache would block until the cache was populated.
+        /// Most implementation will return the same answer every time. Some, like a cache, might
+        /// return null until the cache is fully populated.
         /// </summary>
-        long? GetRowCount(bool lazy = true);
+        long? GetRowCount();
 
         /// <summary>
         /// Get a row cursor. The active column indices are those for which needCol(col) returns true.

src/Microsoft.ML.Data/Data/DataViewUtils.cs

@@ -77,7 +77,7 @@ public static string[] GetTempColumnNames(this ISchema schema, int n, string tag
         /// </summary>
         public static long ComputeRowCount(IDataView view)
         {
-            long? countNullable = view.GetRowCount(lazy: false);
+            long? countNullable = view.GetRowCount();
             if (countNullable != null)
                 return countNullable.Value;
             long count = 0;

src/Microsoft.ML.Data/Data/RowCursorUtils.cs

@@ -541,7 +541,7 @@ public IRowCursor[] GetRowCursorSet(out IRowCursorConsolidator consolidator, Fun
                 return new IRowCursor[] { GetRowCursor(needCol, rand) };
             }
 
-            public long? GetRowCount(bool lazy = true)
+            public long? GetRowCount()
             {
                 return 1;
             }

src/Microsoft.ML.Data/DataLoadSave/Binary/BinaryLoader.cs

@@ -761,7 +761,7 @@ public void GetMetadata<TValue>(string kind, int col, ref TValue value)
 
         private long RowCount { get { return _header.RowCount; } }
 
-        public long? GetRowCount(bool lazy = true) { return RowCount; }
+        public long? GetRowCount() { return RowCount; }
 
         public bool CanShuffle { get { return true; } }
 

src/Microsoft.ML.Data/DataLoadSave/CompositeDataLoader.cs

@@ -557,9 +557,9 @@ private static string GenerateTag(int index)
             return string.Format("xf{0:00}", index);
         }
 
-        public long? GetRowCount(bool lazy = true)
+        public long? GetRowCount()
         {
-            return View.GetRowCount(lazy);
+            return View.GetRowCount();
         }
 
         public bool CanShuffle => View.CanShuffle;

src/Microsoft.ML.Data/DataLoadSave/PartitionedFileLoader.cs

@@ -287,7 +287,7 @@ public void Save(ModelSaveContext ctx)
 
         public Schema Schema { get; }
 
-        public long? GetRowCount(bool lazy = true)
+        public long? GetRowCount()
         {
             return null;
         }

src/Microsoft.ML.Data/DataLoadSave/Text/TextLoader.cs

@@ -1352,7 +1352,7 @@ public BoundLoader(TextLoader reader, IMultiStreamSource files)
                 _files = files;
             }
 
-            public long? GetRowCount(bool lazy = true)
+            public long? GetRowCount()
             {
                 // We don't know how many rows there are.
                 // REVIEW: Should we try to support RowCount?

src/Microsoft.ML.Data/DataLoadSave/Text/TextSaver.cs

@@ -420,7 +420,7 @@ private void WriteDataCore(IChannel ch, TextWriter writer, IDataView data,
                 if (_outputSchema)
                     WriteSchemaAsComment(writer, header);
 
-                double rowCount = data.GetRowCount(true) ?? double.NaN;
+                double rowCount = data.GetRowCount() ?? double.NaN;
                 using (var pch = !_silent ? _host.StartProgressChannel("TextSaver: saving data") : null)
                 {
                     long stateCount = 0;