FillNullByStrategy / Manipulation Layer

Fill null values in columns using predefined strategies. Similar to pandas fillna() or R's na.fill(). This operation handles missing data through various methods, each suited for different analytical needs.

Common applications:

Time series gap filling
Sensor data cleanup
Survey response completion
Financial data preprocessing
Machine learning dataset preparation

Multiple columns can be processed simultaneously, each with its own filling strategy, making it efficient for bulk data cleaning operations.

Table

Transforms

[, ...]

Defines a single column fill operation with its strategy. Multiple transforms can be applied in parallel, each potentially using different strategies based on the data characteristics and analysis requirements.

SelectFill

column

The column containing null values to be filled. The column's data type must be compatible with the chosen strategy (e.g., numeric for Mean strategy).

Strategy

enum

Forward

Available strategies for null value replacement, each designed for specific data patterns and analysis requirements.

Forward ~

Propagate last valid value forward (LOCF - Last Observation Carried Forward). Ideal for:

Time series with stable periods
State-based data
Sensor readings with gaps

Backward ~

Fill with next valid value (NOCB - Next Observation Carried Backward). Useful for:

Retroactive data updates
Backward-looking analysis
End-of-period assignments

Mean ~

Fill with column's mean value. Requires numeric data. Appropriate for:

Statistical analysis
Normal distributions
Centered data estimation

Min ~

Fill with column's minimum value. Suitable for:

Conservative estimates
Lower bound analysis
Worst-case scenarios

Max ~

Fill with column's maximum value. Useful for:

Upper bound analysis
Best-case scenarios
Ceiling estimates

Zero ~

Replace nulls with zero. Common in:

Count data
Financial calculations
Binary indicators

One ~

Replace nulls with one. Useful for:

Multiplicative operations
Unit-based calculations
Neutral scaling factors

MinBound ~

Fill with data type's minimum value (e.g., INT_MIN). Used for:

Range-based analysis
System boundaries
Edge case handling

MaxBound ~

Fill with data type's maximum value (e.g., INT_MAX). Applied in:

System limits
Boundary testing
Maximum range scenarios

AsColumn

name

Name for the new column. If not provided, the system generates a unique name. If AsColumn matches an existing column, the existing column is replaced. The name should follow valid column naming conventions.