apache/iceberg · error · java.lang.UnsupportedOperationException

Spark does not support geography edge algorithm:

Error message

Spark does not support geography edge algorithm: 

What it means

Iceberg geography columns carry an edge interpolation algorithm, but Spark's geography support only recognizes the spherical algorithm. convertAlgorithm throws when a geography column specifies any other edge algorithm.

Source

Thrown at spark/v4.2/spark/src/main/java/org/apache/iceberg/spark/TypeToSparkType.java:192

    // IllegalArgumentException).
    return GeometryType$.MODULE$.apply(geometry.crs());
  }

  private DataType geographyType(Types.GeographyType geography) {
    // The spec requires a geography CRS to be geographic; Spark recognizes only OGC:CRS84, so any
    // other CRS throws SparkIllegalArgumentException (an IllegalArgumentException).
    return GeographyType$.MODULE$.apply(geography.crs(), convertAlgorithm(geography.algorithm()));
  }

  // Translates Iceberg's edge-interpolation algorithm to Spark's. Spark supports only the spherical
  // algorithm (the Iceberg default); every other algorithm is rejected with a clear error. Shared
  // with PruneColumnsWithoutReordering, which compares algorithms across the two type systems.
  static EdgeInterpolationAlgorithm convertAlgorithm(EdgeAlgorithm algorithm) {
    switch (algorithm) {
      case SPHERICAL:
        return SPARK_SPHERICAL;
      default:
        throw new UnsupportedOperationException(
            "Spark does not support geography edge algorithm: " + algorithm);
    }
  }

  private Metadata fieldMetadata(int fieldId) {
    if (MetadataColumns.metadataFieldIds().contains(fieldId)) {
      return new MetadataBuilder().putBoolean(METADATA_COL_ATTR_KEY, true).build();
    }

    return Metadata.empty();
  }
}

View on GitHub (pinned to 86d9c8fc54)

Solutions

1. Rewrite the geography column with edge algorithm spherical (or use geometry instead of geography)
2. Skip/exclude the geography column from Spark queries
3. Upgrade to an Iceberg/Spark version supporting the algorithm

Example fix

// before
Types.GeographyType.get("srid", EdgeAlgorithm.VINCENTY);
// after
Types.GeographyType.get("srid", EdgeAlgorithm.SPHERICAL);

table.schema().columns().stream()
  .filter(c -> c.type() instanceof Types.GeographyType)
  .map(c -> (Types.GeographyType) c.type())
  .forEach(g -> Preconditions.checkArgument(
      g.algorithm() == null || g.algorithm().name().equals("SPHERICAL"),
      "Non-spherical geography not supported in Spark: %s", g));

try { df = spark.read().load("iceberg_geo_table"); }
catch (UnsupportedOperationException e) {
  if (e.getMessage().contains("edge algorithm")) { /* rewrite column or exclude */ }
  else throw e;
}

When it happens

Trigger: Reading a table whose schema has a geography column with an edge algorithm other than spherical (e.g. vincenty or a future spec algorithm) through the Spark type converter.

Common situations: Tables written by geospatial-enabled engines using non-spherical edges, later queried in Spark; spec evolution introducing new algorithms ahead of Spark support.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

• Cannot convert Spark geography with mixed SRID to Iceberg
• Iceberg does not support Spark geography edge algorithm:
• Cannot convert Spark geography with mixed SRID to Iceberg
• Not a supported type: type
• Spark does not support time fields

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/fdc4f56c2c4d0a24. Report an issue: GitHub.