# Import

Dolt supports three modes of import:

1. [`LOAD DATA INFILE`](https://dev.mysql.com/doc/refman/8.0/en/load-data.html) SQL-server command.
2. [`dolt table import`](/cli-reference/cli.md#dolt-table-import) CLI command.
3. `dolt sql < import.sql` batch script import.

We recommend (1) > (2) > (3) for large import performance. `dolt table import` is the most convenient and only slightly slower than `LOAD DATA INFILE`. Refer to the [import tutorial blog](https://www.dolthub.com/blog/2022-11-21-import-perf/) for a walkthrough of the different techniques.

## Import comparison

Each result row reports the runtime for a MySQL import of a certain table schema (`test_name` and `detail`), row count (`row_cnt`), and sorting (`sorting`). `sql_mult` and `cli_mult` distinguish the Dolt SQL (sql-server `LOAD DATA`) and CLI (`dolt table import`) latencies as a multiple of the MySQL latency for the same test conditions. All MySQL tests report the `LOAD DATA` latency.

We are about 2x slower than MySQL for most import conditions. We are slightly slower importing blobs, reflecting how Dolt chunks blobs individually as prolly trees rather than a single byte array. Both Dolt and MySQL are less efficient importing sql scripts with standalone INSERT rows compared to batched insert script (we have a tool [here](https://github.com/dolthub/insert-batcher) to batch INSERT scripts).

| test\_name      | detail       | row\_cnt | sorted | mysql\_time | sql\_mult | cli\_mult |
| --------------- | ------------ | -------- | ------ | ----------- | --------- | --------- |
| batching        | LOAD DATA    | 10000    | 1      | 0.08        | 0.88      |           |
| batching        | batch sql    | 10000    | 1      | 0.13        | 1.62      |           |
| batching        | by line sql  | 10000    | 1      | 0.13        | 1.54      |           |
| blob            | 1 blob       | 200000   | 1      | 1.29        | 3.52      | 3.78      |
| blob            | 2 blobs      | 200000   | 1      | 1.29        | 4.85      | 4.92      |
| blob            | no blob      | 200000   | 1      | 1.3         | 1.58      | 1.67      |
| col type        | datetime     | 200000   | 1      | 1.21        | 2.01      | 2.14      |
| col type        | varchar      | 200000   | 1      | 1           | 2.26      | 2.44      |
| config width    | 2 cols       | 200000   | 1      | 1.13        | 1.51      | 1.55      |
| config width    | 32 cols      | 200000   | 1      | 2.61        | 1.78      | 2.62      |
| config width    | 8 cols       | 200000   | 1      | 1.39        | 1.71      | 1.95      |
| pk type         | float        | 200000   | 1      | 1.28        | 1.37      | 1.38      |
| pk type         | int          | 200000   | 1      | 1.15        | 1.47      | 1.57      |
| pk type         | varchar      | 200000   | 1      | 2.22        | 1.7       | 1.92      |
| row count       | 1.6mm        | 1600000  | 1      | 8.02        | 1.78      | 1.85      |
| row count       | 400k         | 400000   | 1      | 2.06        | 1.66      | 1.68      |
| row count       | 800k         | 800000   | 1      | 4.1         | 1.75      | 1.79      |
| secondary index | four index   | 200000   | 1      | 5.86        | 1.29      | 1.39      |
| secondary index | no secondary | 200000   | 1      | 1.34        | 1.46      | 1.6       |
| secondary index | one index    | 200000   | 1      | 1.73        | 1.6       | 1.7       |
| secondary index | two index    | 200000   | 1      | 3.19        | 1.37      | 1.46      |
| sorting         | shuffled 1mm | 1000000  | 0      | 8.01        | 1.86      | 1.92      |
| sorting         | sorted 1mm   | 1000000  | 1      | 8.57        | 1.71      | 1.8       |


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.dolthub.com/sql-reference/benchmarks/import.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.