What causes "Failure parsing TableRow" in apache/beam?

Inspect the chained exception to identify the malformed row and field causing conversion failure.

Failure parsing TableRow

Error message

Failure parsing TableRow

What it means

While emitting successful rows after a Storage Write API append, Beam converts each serialized proto row back into a TableRow (via protoToTableRow). If conversion of any individual row throws, that row is logged with this warning and skipped rather than failing the whole batch — other rows still produce output.

Solutions

Inspect the chained exception to identify the malformed row and field causing conversion failure.
Ensure the descriptor matches the schema used when rows were serialized (avoid mid-stream schema changes, or use schema-consistent destinations).
Configure failed-row routing (withFailedRows/withErrorFn) so bad rows go to an error output instead of being dropped by this path.
Upgrade Beam — protoToTableRow conversion bugs are fixed periodically.

Defensive patterns

Strategy: try-catch

Validate before calling

// Confirm serialized rows decode with the current descriptor:
try { DynamicMessage.parseFrom(descriptor, rowBytes); } catch (Exception e) { /* route to dead-letter */ }

Type guard

boolean rowParses(Descriptor d, ByteString bytes) { try { DynamicMessage.parseFrom(d, bytes); return true; } catch (Exception e) { return false; } }

Try / catch

try { row = protoToTableRow(DynamicMessage.parseFrom(descriptor, rowBytes)); } catch (Exception e) { LOG.warn("Failure parsing TableRow", e); failedRowsReceiver.output(...); }

Prevention

Route failures with withFailedRows so malformed rows are captured, not dropped.
Avoid schema changes mid-stream that invalidate serialized rows.
Log row bytes and descriptor version when conversion fails to aid diagnosis.

When it happens

Trigger: flush() iterates over successfully appended serialized rows and calls protoToTableRow on DynamicMessage.parseFrom(descriptor, rowBytes); any exception (malformed row bytes, conversion bug, unexpected field values) triggers this per-row warning.

Common situations: Rows written by a different schema version than the descriptor used for parsing; corrupt or truncated serialized rows; custom proto-to-row conversion edge cases (e.g. unusual timestamp/numeric encodings).

Understand the failure class

Background: "cannot parse invalid wire-format data", "cannot unmarshal", "failed unmarshalling": protobuf unmarshal errors explained — this error's family across 10 libraries.

Related errors

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/da47b70d29d20385. Report an issue: GitHub.

Appendix: source

Thrown at sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/StorageApiWriteUnshardedRecords.java:913

                  LOG.warn(
                      "Failure getting proto descriptor. Successful output will not be produced.",
                      e);
                }
                if (descriptor != null) {
                  for (int i = 0; i < c.protoRows.getSerializedRowsCount(); ++i) {
                    ByteString rowBytes = c.protoRows.getSerializedRowsList().get(i);
                    try {
                      TableRow row =
                          TableRowToStorageApiProto.tableRowFromMessage(
                              Preconditions.checkStateNotNull(appendClientInfo)
                                  .getSchemaInformation(),
                              DynamicMessage.parseFrom(descriptor, rowBytes),
                              true,
                              successfulRowsPredicate);
                      org.joda.time.Instant timestamp = c.timestamps.get(i);
                      successfulRowsReceiver.outputWithTimestamp(row, timestamp);
                    } catch (Exception e) {
                      LOG.warn("Failure parsing TableRow", e);
                    }
                  }
                }
              }
            },
            appendRowsContext);
        maybeTickleCache();
        return inserts.getSerializedRowsCount();
      }

      String retrieveErrorDetails(Iterable<AppendRowsContext> failedContext) {
        return StreamSupport.stream(failedContext.spliterator(), false)
            .<@Nullable Throwable>map(AppendRowsContext::getError)
            .filter(Objects::nonNull)
            .map(
                thrw ->
                    Preconditions.checkStateNotNull(thrw).toString()
                        + "\n"

View on GitHub (pinned to 12126d8942)