ingestr benchmark results

Duration vs. rows

Median time per tool across data sizes — log duration scale, lower is faster.

ingestr v1ingestr v0SlingdltAirbyteSpark

Rows

Source

Dest

1K rows

Postgres → Postgres

ingestr v1

0.19s

Sling

0.45s2.4×

dlt

1.56s8.2×

ingestr v0

1.94s10.1×

Spark

6.21s32.4×

Airbyte

14.21s74.2×

Postgres → DuckDB

ingestr v1

0.28s

Sling

1.60s5.8×

dlt

1.71s6.2×

ingestr v0

1.99s7.2×

Spark

7.70s27.8×

DuckDB → DuckDB

ingestr v1

0.25s

Sling

1.77s7.2×

dlt

1.93s7.9×

ingestr v0

2.12s8.6×

Spark

7.42s30.3×

DuckDB → Postgres

ingestr v1

0.25s

Sling

0.63s2.5×

dlt

1.83s7.2×

ingestr v0

1.98s7.8×

Spark

6.98s27.6×

MySQL → Postgres

ingestr v1

0.19s

Sling

0.54s2.9×

dlt

1.49s7.9×

ingestr v0

1.86s9.9×

Spark

6.73s35.7×

Airbyte

18.35s97.3×

MongoDB → Postgres

ingestr v1

0.29s

Sling

0.69s2.4×

dlt

2.20s7.7×

ingestr v0

2.59s9.1×

Airbyte

13.83s48.4×

MongoDB → DuckDB

ingestr v1

0.41s

dlt

1.70s4.1×

Sling

1.89s4.5×

ingestr v0

2.63s6.3×

DuckDB → Snowflake

ingestr v1

3.71s

Sling

5.18s1.4×

dlt

7.97s2.2×

Postgres → SQL Server

ingestr v1

0.22s

Sling

1.10s5.1×

dlt

1.88s8.6×

ingestr v0

2.22s10.2×

Spark

6.62s30.4×

MySQL → SQL Server

ingestr v1

1.08s

Sling

1.16s1.1×

dlt

1.83s1.7×

ingestr v0

2.23s2.1×

Spark

6.71s6.2×

MongoDB → SQL Server

ingestr v1

1.63s

Sling

1.80s1.1×

dlt

2.69s1.6×

DuckDB → SQL Server

ingestr v1

0.28s

Sling

1.27s4.5×

dlt

2.08s7.4×

ingestr v0

2.36s8.3×

Spark

6.89s24.3×

10K rows

Postgres → Postgres

ingestr v1

0.37s

Sling

0.66s1.8×

dlt

1.93s5.2×

ingestr v0

2.24s6.0×

Spark

6.40s17.2×

Airbyte

15.85s42.7×

Postgres → DuckDB

ingestr v1

0.40s

dlt

1.90s4.7×

Sling

1.98s4.9×

ingestr v0

2.21s5.5×

Spark

11.27s28.0×

DuckDB → DuckDB

ingestr v1

0.38s

dlt

2.04s5.3×

Sling

2.26s5.9×

ingestr v0

2.26s5.9×

Spark

10.87s28.2×

DuckDB → Postgres

ingestr v1

0.37s

Sling

1.09s3.0×

dlt

2.11s5.7×

ingestr v0

2.28s6.2×

Spark

6.95s18.9×

MySQL → Postgres

ingestr v1

0.37s

Sling

0.81s2.2×

dlt

1.90s5.1×

ingestr v0

2.33s6.2×

Spark

6.63s17.7×

Airbyte

19.83s53.1×

MongoDB → Postgres

ingestr v1

0.97s

Sling

1.77s1.8×

dlt

5.36s5.5×

ingestr v0

5.64s5.8×

Airbyte

15.37s15.9×

MongoDB → DuckDB

ingestr v1

1.07s

dlt

2.24s2.1×

Sling

3.17s3.0×

ingestr v0

4.59s4.3×

Postgres → SQL Server

ingestr v1

0.50s

dlt

5.03s10.0×

ingestr v0

5.43s10.8×

Sling

6.35s12.6×

Spark

6.83s13.6×

MySQL → SQL Server

ingestr v1

3.00s

dlt

4.87s1.6×

ingestr v0

5.38s1.8×

Sling

6.47s2.2×

Spark

6.67s2.2×

MongoDB → SQL Server

ingestr v1

3.82s

dlt

10.83s2.8×

Sling

12.16s3.2×

DuckDB → SQL Server

ingestr v1

0.51s

dlt

5.20s10.3×

ingestr v0

5.42s10.7×

Sling

6.76s13.4×

Spark

6.88s13.6×

100K rows

Postgres → Postgres

ingestr v1

1.11s

Sling

2.51s2.3×

dlt

5.18s4.7×

ingestr v0

5.71s5.1×

Spark

7.18s6.5×

Airbyte

36.40s32.8×

Postgres → DuckDB

ingestr v1

1.27s

dlt

3.75s3.0×

ingestr v0

4.17s3.3×

Sling

5.42s4.3×

Spark

46.04s36.2×

DuckDB → DuckDB

ingestr v1

1.15s

dlt

3.13s2.7×

ingestr v0

3.34s2.9×

Sling

6.57s5.7×

Spark

45.47s39.5×

DuckDB → Postgres

ingestr v1

0.94s

dlt

4.49s4.8×

ingestr v0

4.72s5.0×

Sling

5.40s5.7×

Spark

7.53s8.0×

MySQL → Postgres

ingestr v1

1.45s

Sling

3.29s2.3×

dlt

5.26s3.6×

ingestr v0

6.64s4.6×

Spark

7.25s5.0×

Airbyte

39.56s27.3×

MongoDB → Postgres

ingestr v1

6.27s

Sling

10.09s1.6×

Airbyte

31.46s5.0×

ingestr v0

36.30s5.8×

dlt

39.65s6.3×

MongoDB → DuckDB

ingestr v1

6.18s

dlt

6.81s1.1×

Sling

13.05s2.1×

ingestr v0

28.29s4.6×

Postgres → Snowflake

ingestr v1

5.71s

Sling

8.20s1.4×

dlt

10.73s1.9×

ingestr v0

11.87s2.1×

MySQL → Snowflake

ingestr v1

4.24s

Sling

9.11s2.1×

dlt

11.52s2.7×

ingestr v0

12.17s2.9×

MongoDB → Snowflake

ingestr v1

5.58s

Sling

12.58s2.3×

dlt

18.48s3.3×

ingestr v0

18.79s3.4×

DuckDB → Snowflake

ingestr v1

4.19s

dlt

10.10s2.4×

Sling

10.70s2.6×

ingestr v0

10.84s2.6×

Postgres → SQL Server

ingestr v1

1.45s

Spark

8.85s6.1×

dlt

37.91s26.2×

ingestr v0

37.92s26.2×

Sling

54.86s37.9×

MySQL → SQL Server

ingestr v1

3.54s

Spark

8.62s2.4×

dlt

37.50s10.6×

ingestr v0

39.33s11.1×

Sling

48.38s13.7×

MongoDB → SQL Server

ingestr v1

7.60s

dlt

92.36s12.2×

Sling

112.8s14.8×

DuckDB → SQL Server

ingestr v1

1.19s

Spark

9.06s7.6×

ingestr v0

37.38s31.4×

dlt

37.52s31.5×

Sling

50.44s42.3×

1M rows

Postgres → Postgres

ingestr v1

6.90s

Spark

12.40s1.8×

Sling

20.83s3.0×

dlt

36.27s5.3×

ingestr v0

41.09s6.0×

Airbyte

240.6s34.9×

Postgres → DuckDB

ingestr v1

8.20s

dlt

18.91s2.3×

ingestr v0

21.25s2.6×

Sling

38.99s4.8×

Spark

386.1s47.1×

DuckDB → DuckDB

ingestr v1

7.63s

dlt

12.38s1.6×

ingestr v0

12.45s1.6×

Sling

47.61s6.2×

Spark

384.7s50.4×

DuckDB → Postgres

ingestr v1

4.98s

Spark

13.35s2.7×

dlt

27.75s5.6×

ingestr v0

30.48s6.1×

Sling

47.15s9.5×

MySQL → Postgres

ingestr v1

11.43s

Spark

13.04s1.1×

Sling

27.60s2.4×

dlt

38.32s3.4×

ingestr v0

52.59s4.6×

Airbyte

239.2s20.9×

MongoDB → Postgres

ingestr v1

59.59s

Sling

90.69s1.5×

Airbyte

183.0s3.1×

ingestr v0

319.0s5.4×

dlt

382.5s6.4×

MongoDB → DuckDB

dlt

47.94s

ingestr v1

54.46s1.1×

Sling

110.6s2.3×

ingestr v0

261.6s5.5×

Postgres → Snowflake

ingestr v1

11.59s

dlt

30.62s2.6×

ingestr v0

31.78s2.7×

Sling

32.70s2.8×

MySQL → Snowflake

ingestr v1

10.18s

dlt

38.23s3.8×

Sling

40.85s4.0×

ingestr v0

42.54s4.2×

MongoDB → Snowflake

ingestr v1

23.48s

Sling

70.85s3.0×

ingestr v0

93.07s4.0×

dlt

101.4s4.3×

DuckDB → Snowflake

ingestr v1

8.88s

ingestr v0

22.69s2.6×

dlt

23.04s2.6×

Sling

58.59s6.6×

Postgres → SQL Server

ingestr v1

8.58s

Spark

21.84s2.5×

ingestr v0

91.88s10.7×

dlt

377.8s44.0×

Sling

629.9s73.4×

MySQL → SQL Server

ingestr v1

13.57s

Spark

26.98s2.0×

ingestr v0

104.3s7.7×

dlt

372.6s27.5×

Sling

430.9s31.7×

MongoDB → SQL Server

ingestr v1

49.61s

dlt

930.9s18.8×

Sling

1,642.80s33.1×

DuckDB → SQL Server

ingestr v1

5.72s

Spark

25.19s4.4×

ingestr v0

81.45s14.2×

dlt

372.6s65.1×

Sling

453.1s79.1×

10M rows

Postgres → Postgres

ingestr v1

64.50s

Spark

65.34s1.0×

Sling

205.5s3.2×

dlt

344.2s5.3×

ingestr v0

384.0s6.0×

Postgres → DuckDB

ingestr v1

73.26s

dlt

158.1s2.2×

ingestr v0

191.5s2.6×

Sling

373.4s5.1×

Spark

3,887.86s53.1×

DuckDB → DuckDB

ingestr v1

71.97s

dlt

92.74s1.3×

ingestr v0

102.6s1.4×

Sling

454.9s6.3×

Spark

3,853.76s53.5×

DuckDB → Postgres

ingestr v1

45.21s

Spark

70.56s1.6×

dlt

257.6s5.7×

ingestr v0

279.3s6.2×

Sling

466.5s10.3×

MySQL → Postgres

Spark

70.70s

ingestr v1

111.4s1.6×

Sling

274.1s3.9×

dlt

368.8s5.2×

ingestr v0

513.5s7.3×

MongoDB → Postgres

ingestr v1

597.4s

Sling

885.8s1.5×

ingestr v0

3,204.22s5.4×

dlt

4,043.57s6.8×

MongoDB → DuckDB

dlt

451.1s

ingestr v1

529.5s1.2×

Sling

1,086.43s2.4×

ingestr v0

2,624.34s5.8×

Postgres → SQL Server

ingestr v1

80.62s

Spark

216.1s2.7×

ingestr v0

738.9s9.2×

dlt

3,718.56s46.1×

MySQL → SQL Server

ingestr v1

93.58s

Spark

210.6s2.3×

ingestr v0

877.9s9.4×

dlt

3,693.68s39.5×

DuckDB → SQL Server

ingestr v1

53.16s

Spark

182.6s3.4×

ingestr v0

629.2s11.8×

dlt

3,637.23s68.4×

Methodology

Each tool runs the same source-to-destination copy for a fixed number of rows. Times are wall-clock, measured with hyperfine, with a warmup run discarded and the mean of the timed runs reported. The multiplier shows how many times slower a tool is than the fastest one in that chart.

All tools were run on the same machine with identical source data and warehouse targets. Numbers are reproducible from the raw benchmark dumps — see the benchmark code. Runs are refreshed as engines change — the most recent run wins for each source / destination / row-count combination.

Benchmarks were executed on a Hetzner Cloud CCX63 instance (48 dedicated AMD EPYC-Milan vCPUs, 192 GB RAM, NVMe SSD), running Ubuntu 24.04.4 LTS.