Want to import data from JSON Lines (JSONL) to SQL Server via Command Line Interface (CLI) on Windows?
Using mssql-cmd, a Command Line Interface (CLI) client for SQL Server, you can import data from JSON Lines to SQL Server easily and fast. Direct Interaction or Scripting and Automation.
- No need to install SQL Server client library.
- Convenient to deploy to cloud. No need to run installation file, just uncompress and run, immediately.
- Commands are entered into a terminal or command prompt, accessed from local or remote (like SSH).
- Support Windows, Linux, and MacOS.
Download mssql-cmd Windows version 
JSON Lines text format, also called newline-delimited JSON (NDJSON), or Line-delimited JSON (LDJSON). JSON Lines files may be saved with the file extension .jsonl, *.ndjson, *.ldjson.
Import data from JSON Lines to SQL Server via command line on Windows
1. In interactive mode:
mssql-cmd> load table=media_messages
> datafile=f:\media_messages.json
> filetype=json;
2. In batch mode:
mssql-cmd dsn="pan:withdata@192.168.0.102:3306/test" task=load table=media_messages datafile=f:\media_messages.json filetype=json quit=y
“loadtype=replace” will delete all data in table before inserting.
See parameters detail.
More about mssql-cmd – Command Line Interface (CLI) tool for SQL Server
Download mssql-cmd Windows version
JSON → SQL Server reference
Type mapping
| Value kind | in JSON | in SQL Server |
|---|---|---|
| Int | 1234 (number) | int |
| Bigint | 9007199254740993 | bigint |
| Decimal | 1234.56 (number) | decimal(p,s) |
| Float | 1.234e10 (number) | float |
| Text | "text" (string) | nvarchar(n) |
| Long text | "multi-line\ntext" (escaped newline inside a string) | nvarchar(max) |
| Date | "2026-09-18" (string) | date |
| Datetime | "2026-09-18T14:30:00Z" (string) | datetime2 |
| Binary | "base64:..." (string) | varbinary(max) |
| Boolean | true / false (boolean) | bit |
| JSON | nested object or array, kept as structure | nvarchar(max) |
Dialect differences
JSON is a file format and SQL Server is a database server, so the target has to be told the types the file never carried – the rows below are where the two sides diverge.
| Property | in JSON | in SQL Server |
|---|---|---|
| port | — | 1433 |
| identifiers | Double quotes; backslash escapes | [square brackets] |
| row limit | — | SELECT TOP n |
| dates | No date type: dates arrive as strings and the reader must p... | datetime, datetime2, date, time |
| long text | — | nvarchar(max) |
| binary | Not supported natively - base64 inside a string | varbinary(max) |
| separator | Braces and commas | — |
| encoding | UTF-8 by definition | — |
| typing | string, number, boolean, null, object, array - no date type | — |
| shape | Nested objects and arrays | Tables of rows and columns |
| ceiling | Whole document is usually parsed into memory | — |
Command line
| Side | Command |
|---|---|
| SQL Server | bcp |
| SQL Server (load) | bcp, or BULK INSERT |
Convert your data online, free
Free in-browser converters for CSV, Excel, JSON, XML and SQL: Withdata Online Data Conversion. Large files, batch conversion or scheduling? See the offline desktop apps, or browse all data conversion tools.
