> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/sdgm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/sdgm/_mcp/server.

# How does Kumo handle missing values in my dataset?

> Configure missing-value handling through column encoders

Column encoders determine how Kumo Engine handles missing values. Numerical encoders support mean imputation, zero imputation, or a separate missing-value representation. Other encoders support different strategies.

Set defaults by semantic type with `plan.column_processing.na_strategy`, or configure individual columns through `encoder_overrides`. See [Missing Values](/kumo-engine/user-code-guide/set-up-a-graph/encoder-selection#missing-values) for supported values and precedence.

Convert placeholders such as `-1` or `"N/A"` to nulls when they represent missing data. Check the source reader's parsing rules; a sentinel value is not necessarily read as null.