apache/iceberg · error · IcebergAnalysisException

ALTER VIEW AS is not supported. Use CREATE OR REPLACE VIEW…

Error message

ALTER VIEW <viewName> AS is not supported. Use CREATE OR REPLACE VIEW instead

What it means

Iceberg's view support in Spark 4.0 does not implement ALTER VIEW ... AS (redefining a view in place). When CheckViews sees AlterViewAs applied to a resolved Iceberg V2 view, it throws this IcebergAnalysisException and directs users to CREATE OR REPLACE VIEW, which Iceberg supports for full view redefinition.

Solutions

Rewrite the statement as CREATE OR REPLACE VIEW catalog.db.v AS SELECT ...
Drop and recreate the view if replace semantics are unacceptable
Use ALTER VIEW only for operations Iceberg supports (e.g. rename is via ALTER VIEW RENAME TO)

Example fix

// before
ALTER VIEW my_view AS SELECT id, name FROM t
// after
CREATE OR REPLACE VIEW my_view AS SELECT id, name FROM t

Defensive patterns

Strategy: validation

Validate before calling

// static check: replace any occurrence in your SQL pipeline
val fixed = sql.replaceAll(?i"ALTER\s+VIEW\s+(\S+)\s+AS", "CREATE OR REPLACE VIEW $1 AS")

Try / catch

try {
  spark.sql(stmt)
} catch {
  case e: Exception if e.getMessage.contains("ALTER VIEW") && e.getMessage.contains("CREATE OR REPLACE VIEW") =>
    val rewritten = rewriteAlterViewAs(stmt)
    spark.sql(rewritten)
}

Prevention

Use CREATE OR REPLACE VIEW as the standard redefinition idiom for Iceberg views
Lint SQL scripts for ALTER VIEW ... AS against Iceberg catalogs
Review migration scripts from the built-in session catalog for unsupported DDL

When it happens

Trigger: Running ALTER VIEW catalog.db.v AS SELECT ... against an Iceberg view (a ResolvedV2View) via the Iceberg Spark extensions.

Common situations: Scripts written for the built-in Spark session catalog ported to Iceberg catalogs where ALTER VIEW AS is unavailable; schema-evolution workflows attempting to redefine a view in place.

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

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/936cafc980c2558f. Report an issue: GitHub.

Appendix: source

Thrown at spark/v4.0/spark-extensions/src/main/scala/org/apache/spark/sql/catalyst/analysis/CheckViews.scala:64

            _,
            _,
            _,
            _,
            replace,
            _,
            _) =>
        verifyColumnCount(resolvedIdent, columnAliases, query)
        SchemaUtils.checkColumnNameDuplication(
          query.schema.fieldNames.toIndexedSeq,
          SQLConf.get.resolver)
        if (replace) {
          val viewIdent: Seq[String] =
            resolvedIdent.catalog.name() +: resolvedIdent.identifier.asMultipartIdentifier
          checkCyclicViewReference(viewIdent, query, Seq(viewIdent))
        }

      case AlterViewAs(ResolvedV2View(_, _), _, _) =>
        throw new IcebergAnalysisException(
          "ALTER VIEW <viewName> AS is not supported. Use CREATE OR REPLACE VIEW instead")

      case _ => // OK
    }
  }

  private def verifyColumnCount(
      ident: ResolvedIdentifier,
      columns: Seq[String],
      query: LogicalPlan): Unit = {
    if (columns.nonEmpty) {
      if (columns.length > query.output.length) {
        throw new AnalysisException(
          errorClass = "CREATE_VIEW_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS",
          messageParameters = Map(
            "viewName" -> String.format("%s.%s", ident.catalog.name(), ident.identifier),
            "viewColumns" -> columns.mkString(", "),
            "dataColumns" -> query.output.map(c => c.name).mkString(", ")))

View on GitHub (pinned to 86d9c8fc54)