egalpin commented on code in PR #10234: URL: https://github.com/apache/pinot/pull/10234#discussion_r1120988051
########## pinot-segment-local/src/main/java/org/apache/pinot/segment/local/indexsegment/mutable/MutableSegmentImpl.java: ########## @@ -558,15 +560,40 @@ private RecordInfo getRecordInfo(GenericRow row, int docId) { PrimaryKey primaryKey = row.getPrimaryKey(_schema.getPrimaryKeyColumns()); if (isUpsertEnabled()) { - Object upsertComparisonValue = row.getValue(_upsertComparisonColumn); - Preconditions.checkState(upsertComparisonValue instanceof Comparable, - "Upsert comparison column: %s must be comparable", _upsertComparisonColumn); - return new RecordInfo(primaryKey, docId, (Comparable) upsertComparisonValue); + if (_upsertComparisonColumns.size() > 1) { + return multiComparisonRecordInfo(primaryKey, docId, row); + } + Comparable comparisonValue = (Comparable) row.getValue(_upsertComparisonColumns.get(0)); + return new RecordInfo(primaryKey, docId, comparisonValue); } return new RecordInfo(primaryKey, docId, null); } + private RecordInfo multiComparisonRecordInfo(PrimaryKey primaryKey, int docId, GenericRow row) { + int numComparisonColumns = _upsertComparisonColumns.size(); + Comparable[] comparisonValues = new Comparable[numComparisonColumns]; + + int numNonNull = 0; + for (int i = 0; i < numComparisonColumns; i++) { + String columnName = _upsertComparisonColumns.get(i); + + if (!row.isNullValue(columnName)) { + // Inbound records may only have exactly 1 non-null value in one of the comparison column i.e. comparison + // columns are mutually exclusive + numNonNull++; + + Object comparisonValue = row.getValue(columnName); + Preconditions.checkState(comparisonValue instanceof Comparable, + "Upsert comparison column: %s must be comparable", columnName); + comparisonValues[i] = (Comparable) comparisonValue; + } + } + Preconditions.checkState(numNonNull == 1 || numNonNull == 0, Review Comment: > For multiple comparison columns case, there should be at least one comparison column not null. If one event does not have any non-null value, we can just drop it (unless there is no event associated with the key) since we don't know which value to compare. In that case, should we make this check strictly `== 1` ? And avoid the need to handle the case of all null Comparables[] within ComparisonColumns? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org