wgtmac commented on code in PR #801:
URL: https://github.com/apache/iceberg-cpp/pull/801#discussion_r3675784396
##########
src/iceberg/inspect/snapshots_table.cc:
##########
@@ -65,4 +68,50 @@ Result<std::unique_ptr<SnapshotsTable>> SnapshotsTable::Make(
return std::unique_ptr<SnapshotsTable>(new SnapshotsTable(std::move(table)));
}
+Result<ArrowArray> SnapshotsTable::Scan(
+ const std::optional<SnapshotSelection>& /*snapshot_selection*/) {
+ ICEBERG_ASSIGN_OR_RAISE(auto builder, ArrowRowBuilder::Make(*schema()));
+
+ for (const auto& snapshot : source_table()->snapshots()) {
+ if (snapshot == nullptr) [[unlikely]] {
+ continue;
+ }
+
+ // column 0: committed_at (timestamptz -> int64 micros)
+ ICEBERG_RETURN_UNEXPECTED(AppendInt(
+ builder.column(0),
std::chrono::duration_cast<std::chrono::microseconds>(
+ snapshot->timestamp_ms.time_since_epoch())
+ .count()));
+
+ // column 1: snapshot_id (long)
+ ICEBERG_RETURN_UNEXPECTED(AppendInt(builder.column(1),
snapshot->snapshot_id));
+
+ // column 2: parent_id (long, optional)
+ if (snapshot->parent_snapshot_id.has_value()) {
+ ICEBERG_RETURN_UNEXPECTED(
+ AppendInt(builder.column(2), *snapshot->parent_snapshot_id));
+ } else {
+ ICEBERG_RETURN_UNEXPECTED(AppendNull(builder.column(2)));
+ }
+
+ // column 3: operation (string, optional)
+ auto op = snapshot->Operation();
+ if (op.has_value()) {
+ ICEBERG_RETURN_UNEXPECTED(AppendString(builder.column(3), *op));
+ } else {
+ ICEBERG_RETURN_UNEXPECTED(AppendNull(builder.column(3)));
+ }
+
+ // column 4: manifest_list (string, optional)
+ ICEBERG_RETURN_UNEXPECTED(AppendString(builder.column(4),
snapshot->manifest_list));
+
+ // column 5: summary (map<string,string>)
+ ICEBERG_RETURN_UNEXPECTED(AppendStringMap(builder.column(5),
snapshot->summary));
Review Comment:
`Snapshot::summary` is not equivalent to Java’s `Snapshot.summary()` here.
In C++, the map still contains the `operation` entry, while Java stores the
operation separately and `SnapshotsTable.snapshotToRow` emits the remaining
summary map. As a result, every normal row duplicates `operation` in both
columns (the copied Java fixture has 10 summary entries, not 11). Please omit
`SnapshotSummaryFields::kOperation` when materializing this map, while
preserving null versus empty-map semantics.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
