SQL Server CREATE TABLE not setting columns NULLABLE by default

[kenny-statsig]

Software versions

• Tedious: 19
• SQL Server: default (no external library)
• Node.js: 23.6.1

Additional Libraries Used and Versions

Table schema

Connection configuration

      server: data.connectionString,
      authentication: {
        type: 'azure-active-directory-service-principal-secret',
        options: {
          clientId: data.clientId,
          clientSecret: data.clientSecret,
          tenantId: data.tenantId,
        },
      }
      options: {
        encrypt: true,
        database: data.database,
        port: 1433,
        rowCollectionOnRequestCompletion: true,
        connectTimeout: 30000,
        requestTimeout: 30000,
      },

Problem description

Executing the following query using tedious creates a table but the resulting table's columns are all NON NULL columns. Executing it directly from Fabric Warehouse successfully defines the columns as NULL-able

IF NOT EXISTS (
    SELECT * FROM
        INFORMATION_SCHEMA.TABLES
    WHERE
        TABLE_SCHEMA = 'dbo'
        AND
            TABLE_NAME = 'pipeline_overview'
)
CREATE TABLE [dbo].[pipeline_overview] (
    ts DATETIME2 (6),
    dag_id VARCHAR (500),
    job_type VARCHAR (500),
    metric_source_id VARCHAR (500),
    assignment_source_id VARCHAR (500),
    job_status VARCHAR (500),
    metrics VARCHAR (MAX),
    dag_state VARCHAR (500),
    dag_type VARCHAR (500),
    experiment_id VARCHAR (500),
    dag_start_ds VARCHAR (500),
    dag_end_ds VARCHAR (500),
    wall_time NUMERIC,
    turbo_mode BIT,
    dag_duration NUMERIC,
    is_scheduled BIT,
    experiment_team VARCHAR (500),
    metric_source_team VARCHAR (500),
    creator VARCHAR (500)
);

Expected behavior

The columns should be NULL-able.
Executing the following query should work

INSERT INTO [dbo].[pipeline_overview] (
    ts,
    job_type,
    metric_source_id,
    assignment_source_id,
    job_status,
    metrics,
    dag_state,
    dag_type,
    experiment_id,
    dag_start_ds,
    dag_end_ds,
    wall_time,
    turbo_mode,
    dag_id,
    dag_duration,
    is_scheduled,
    experiment_team,
    metric_source_team,
    creator
)
VALUES (
    '2025-03-14T18:55:13.557Z',
    'assignment_source_sync',
    NULL,
    NULL,
    'success',
    JSON_ARRAY (
    ),
'success',
'assignment_source',
'6aX4dteZLrnzNd5zU0wy0T',
'2025-03-14',
'2025-03-14',
5892.51879099998,
0,
'6Idu2zWvr7xScQQxr9Z4zx',
0,
0,
NULL,
NULL,
'Kenny Yi'
),

    (
        '2025-03-14T18:55:13.557Z',
        'assignment_source_sync',
        NULL,
        NULL,
        'success',
JSON_ARRAY 
    (
    ),
    'success',
    'assignment_source',
    '6aX4dteZLrnzNd5zU0wy0T',
    '2025-03-14',
    '2025-03-14',
    5872.721833999996,
    0,
    '6Idu2zWvr7xScQQxr9Z4zx',
    0,
    0,
    NULL,
    NULL,
    'Kenny Yi'
    )

Actual behavior

The columns are NOT NULL-able.
Executing the query gives the error below

Error message/stack trace

 Cannot insert the value NULL into column 'metric_source_id', table 'pipeline_overview'; column does not allow nulls. INSERT fails

Any other details that can be helpful

I am on a forked version of the library https://github.com/statsig-io/tedious. Based on the uncommitted PR #1668

[khkiley]

Interesting, what happens when you specify the nullability of the column, i.e. :

CREATE TABLE [dbo].[pipeline_overview] (
    ts DATETIME2 (6) NOT NULL,
    dag_id VARCHAR (500) NOT NULL,
    job_type VARCHAR (500) NULL,
    metric_source_id VARCHAR (500) NULL ,
    ...
)

Have you tried executing the CREATE TABLE statement directly against the database ?

[kenny-statsig]

Executing the same CREATE TABLE query directly against the database makes the columns NULLABLE.
I have worked around this by manually specifying the nullability. But I would expect this not to be necessary.