Virtual (Deedle) | Deedle

Virtual Type

Namespace: Deedle.Virtual
Assembly: Deedle.dll
Base Type: obj

Provides static methods for creating virtual series and virtual frames. Those provide necessary wrapping around `IVirtualVectorSource` values

Static members

Static member	Description
`Virtual.CreateFrame(indexSource, keys, sources)` Full Usage: `Virtual.CreateFrame(indexSource, keys, sources)` Parameters: indexSource : `IVirtualVectorSource<'a>` keys : `'b seq` sources : `IVirtualVectorSource seq` Returns: `Frame<'a, 'b>` Type parameters: 'a, 'b (requires equality and equality)	Create a frame with ordinal index, containing the specified sources as columns. The index source should support lookup (which is used for series lookup, slicing etc.) The value source does not need to implement lookup - mainly `ValueAt`, merging and getting sub-source indexSource : `IVirtualVectorSource<'a>` keys : `'b seq` sources : `IVirtualVectorSource seq` Returns: `Frame<'a, 'b>`
`Virtual.CreateOrdinalFrame(keys, sources)` Full Usage: `Virtual.CreateOrdinalFrame(keys, sources)` Parameters: keys : `'a seq` sources : `IVirtualVectorSource seq` Returns: `Frame<int64, 'a>` Type parameters: 'a (requires equality)	Create a frame with ordinal index, containing the specified sources as columns. keys : `'a seq` sources : `IVirtualVectorSource seq` Returns: `Frame<int64, 'a>`
`Virtual.CreateOrdinalSeries(source)` Full Usage: `Virtual.CreateOrdinalSeries(source)` Parameters: source : `IVirtualVectorSource<'a>` Returns: `Series<int64, 'a>` Type parameters: 'a	Creates a virtual series with ordinal index. The parameter is `IVirtualVectorSource` that specifies how to access values in the series (and is also used to determine the size of the series index) source : `IVirtualVectorSource<'a>` Returns: `Series<int64, 'a>`
`Virtual.CreateSeries(indexSource, valueSource)` Full Usage: `Virtual.CreateSeries(indexSource, valueSource)` Parameters: indexSource : `IVirtualVectorSource<'a>` valueSource : `IVirtualVectorSource<'b>` Returns: `Series<'a, 'b>` Type parameters: 'a, 'b (requires equality)	Create a virtual series with an index and values specified by two `IVirtualVectorSource` values. The index source should support lookup (which is used for series lookup, slicing etc.) The value source does not need to implement lookup - mainly `ValueAt`, merging and getting sub-source indexSource : `IVirtualVectorSource<'a>` valueSource : `IVirtualVectorSource<'b>` Returns: `Series<'a, 'b>`
`Virtual.Describe(frame)` Full Usage: `Virtual.Describe(frame)` Parameters: frame : `Frame<'R, 'C>` Returns: `string` Type parameters: 'R, 'C (requires equality and equality)	Short human-readable summary of row count, row-index kind, and column count. frame : `Frame<'R, 'C>` Returns: `string`
`Virtual.GetRowIndexKind(frame)` Full Usage: `Virtual.GetRowIndexKind(frame)` Parameters: frame : `Frame<'R, 'C>` Returns: `VirtualRowIndexKind` Type parameters: 'R, 'C (requires equality and equality)	Classify how the frame's row index is stored (ordered/ordinal virtual vs linear). frame : `Frame<'R, 'C>` Returns: `VirtualRowIndexKind`
`Virtual.IsVirtualColumn(frame, column)` Full Usage: `Virtual.IsVirtualColumn(frame, column)` Parameters: frame : `Frame<'R, 'C>` column : `'C` Returns: `bool` Type parameters: 'R, 'C (requires equality and equality)	True when the named column uses a virtual addressing scheme. frame : `Frame<'R, 'C>` column : `'C` Returns: `bool`
`Virtual.IsVirtualRowIndex(frame)` Full Usage: `Virtual.IsVirtualRowIndex(frame)` Parameters: frame : `Frame<'R, 'C>` Returns: `bool` Type parameters: 'R, 'C (requires equality and equality)	True when the row index is ordered or ordinal virtual. frame : `Frame<'R, 'C>` Returns: `bool`
`Virtual.MaterializeFloatBatches(frame, batchSize, columns, ?missingPolicy, ?includeRowKeys, ?labelsColumn, ?layout, ?includeMissingMask, ?maxRows, ?order)` Full Usage: `Virtual.MaterializeFloatBatches(frame, batchSize, columns, ?missingPolicy, ?includeRowKeys, ?labelsColumn, ?layout, ?includeMissingMask, ?maxRows, ?order)` Parameters: frame : `Frame<'TRowKey, 'TColumnKey>` - Source frame (virtual or in-memory). batchSize : `int64` - Positive number of rows per batch (last batch may be smaller). columns : `'TColumnKey list` - Column keys to materialize (`float` or `int64`). ?missingPolicy : `FloatMissingPolicy` - How to replace missing cells (default FloatMissingPolicy.NaN). ?includeRowKeys : `bool` - When set, also copy this batch's row keys (may touch the key source). ?labelsColumn : `'TColumnKey` - Optional single label/target column (`float` or `int64`, same missing policy). ?layout : `FloatBatchLayout` - Row-major (default) or column-major flat layout. ?includeMissingMask : `bool` - When true, set MissingMask for feature cells. ?maxRows : `int64` - Optional cap on total rows exported across all batches. ?order : `FloatBatchOrder` - Row order within batches (default sequential). FloatBatchOrder.Shuffled shuffles rows once per enumeration. Returns: `FloatBatch<'TRowKey> seq` Type parameters: 'TRowKey, 'TColumnKey (requires equality and equality)	Materialize selected numeric columns as a lazy sequence of mini-batches. Each batch is produced by slicing the frame and reading only those rows × columns. With FloatBatchOrder.Sequential (default), batches are contiguous address ranges. With random order, rows are permuted once when enumeration starts and each row appears in exactly one batch. Enumerating later batches does not re-read earlier rows or allocate a full-frame feature matrix. frame : `Frame<'TRowKey, 'TColumnKey>` Source frame (virtual or in-memory). batchSize : `int64` Positive number of rows per batch (last batch may be smaller). columns : `'TColumnKey list` Column keys to materialize (`float` or `int64`). ?missingPolicy : `FloatMissingPolicy` How to replace missing cells (default FloatMissingPolicy.NaN). ?includeRowKeys : `bool` When set, also copy this batch's row keys (may touch the key source). ?labelsColumn : `'TColumnKey` Optional single label/target column (`float` or `int64`, same missing policy). ?layout : `FloatBatchLayout` Row-major (default) or column-major flat layout. ?includeMissingMask : `bool` When true, set MissingMask for feature cells. ?maxRows : `int64` Optional cap on total rows exported across all batches. ?order : `FloatBatchOrder` Row order within batches (default sequential). FloatBatchOrder.Shuffled shuffles rows once per enumeration. Returns: `FloatBatch<'TRowKey> seq`
`Virtual.TryGetLookupRange(frame, column)` Full Usage: `Virtual.TryGetLookupRange(frame, column)` Parameters: frame : `Frame<'R, 'C>` column : `'C` Returns: `VirtualColumnLookupRange option` Type parameters: 'R, 'C (requires equality and equality)	LookupRange kind for a column listed in `searchColumns` at load (inferred or explicit). `None` when the column was not configured as searchable or is not virtual. frame : `Frame<'R, 'C>` column : `'C` Returns: `VirtualColumnLookupRange option`
`Virtual.TryGetRowIndexSchemeId(frame)` Full Usage: `Virtual.TryGetRowIndexSchemeId(frame)` Parameters: frame : `Frame<'R, 'C>` Returns: `string option` Type parameters: 'R, 'C (requires equality and equality)	Scheme id from the virtual row-index source, when present (e.g. `"csv-file"`, instrumented test ids). frame : `Frame<'R, 'C>` Returns: `string option`