Log Tables

The Go SDK currently supports both indexed log writes and Arrow-oriented log flows.

For most readers, the simplest way to understand log tables is:

• append a row
• scan or fetch rows back
• use schema-aware decode helpers

Indexed Row Append

The dev harness writes an indexed log row like this:

row, err := client.NewRow(
	logSchema,
	int64(time.Now().UnixNano()%1_000_000),
	int64(101),
	mustDecimal("19.95", 15, 2),
	"created",
)
if err != nil {
	log.Fatal(err)
}

result, err := logClient.AppendIndexedRowAuto(ctx, row)
if err != nil {
	log.Fatal(err)
}

The Auto form is the preferred helper-first path when bucket routing can be derived from table metadata.

Read Back with Limit Scan

limitResult, err := logClient.LimitScan(ctx, nil, 0, 5)
if err != nil {
	log.Fatal(err)
}

rows, err := client.DecodeIndexedLimitScanRows(logSchema, limitResult)
if err != nil {
	log.Fatal(err)
}

Fetch-Based Log Reads

The lower-level fetch path is also available:

fetched, err := table.FetchLog(ctx, -1, 4096, nil, nil, []client.FetchBucketRequest{{
	BucketID:    0,
	FetchOffset: 0,
}})

This is especially relevant for Arrow log batches and projection-aware reads.

Current Guidance

• Prefer row-oriented helpers first.
• Use fetch APIs when you need request-level control or Arrow-specific flows.
• Treat bucket 0 in examples as demo-driven context, not a universal rule for all deployments.