Upgrade to dbt™ 1.10

dbt Core™ 1.10 adds sample mode for faster development builds and, more importantly, starts warning about project code that will become invalid in future versions. Cleaning up these deprecation warnings now is the bulk of the work of preparing for dbt™ v2. This guide is based on the official dbt™ upgrade guide for v1.10; check the dbt Core™ 1.10 changelog for the full detail.

Prerequisites

An Admin role in Paradime, needed to change the workspace dbt™ version.
Your project on dbt™ 1.9, upgraded via the 1.9 guide.

Estimated time: 30 minutes, plus deprecation cleanup.

What can break

Upgrading itself is backward compatible: the new deprecation warnings are warnings, not errors, and fixing them is not a prerequisite for using 1.10. Two setups need attention:

If you run with --warn-error (or --warn-error-options '{"error": "all"}'), the new deprecation warnings get promoted to errors and can fail Bolt runs that passed on 1.9. To keep deprecations as warnings while still promoting everything else, set "warn": ["Deprecations"] in warn_error_options (new in 1.10).

Snowflake column size change (September 2026). Snowflake plans to increase the default column size for string and binary types, and dbt-snowflake below 1.10.6 may fail to build incremental models that combine collated string columns with on_schema_change='sync_all_columns'. Run dbt ls -s config.materialized:incremental,config.on_schema_change:sync_all_columns --resource-type model to check exposure.

What’s new in dbt™ 1.10

The `--sample` flag

The --sample flag, available on dbt run and dbt build, runs dbt™ in sample mode: refs and sources are filtered with time-based sampling so you can validate model outputs without building entire models. This cuts development build times and warehouse cost on large datasets.

The `anchors:` key for standalone YAML anchors

dbt™ 1.10 warns on unexpected top-level keys in properties YAML files. The common legitimate case is standalone YAML anchor definitions; move those under the new top-level anchors: key:

anchors:
  - &id_column_alias
      name: id
      description: This is a unique identifier.
      data_type: int
      data_tests:
        - not_null
        - unique

models:
  - name: my_first_model
    columns:
      - *id_column_alias
      - name: unrelated_column_a
        description: This column is not repeated in other models
  - name: my_second_model
    columns:
      - *id_column_alias

Only fragments defined outside the main YAML structure need to move. See the anchors reference.

Parsing `catalogs.yml`

dbt™ can now parse a catalogs.yml file, the first step toward external catalog support for Iceberg tables with write integrations. You define a catalog with write_integrations (external volume, table format, catalog type) and point producer models at it with catalog_name in the model config. See external catalog support.

New behavior flags

Both introduced in 1.10 and disabled by default; opt in under flags: in dbt_project.yml:

Flag	What it does when enabled
`validate_macro_args`	Warns when argument names or types in macro YAML don’t match the macro’s Jinja definition or use unsupported types. With `--warn-error`, out-of-sync declarations can fail the run at parse time.
`require_all_warnings_handled_by_warn_error`	Routes all warnings through the `--warn-error` / `--warn-error-options` handler for consistent promoting and silencing. With `--warn-error`, previously ignored warnings can start failing builds, so enable it gradually, one project at a time.

Deprecation warnings

Starting in 1.10, dbt™ warns about code that will become invalid in future versions. Fixing these is not required to use 1.10, but it is required before moving to v2. Many are auto-fixable with the dbt-autofix tool, or with the DinoAI upgrade agent. Custom properties and configs must move under meta. dbt™ previously ignored unrecognized properties (including misspellings like desciption:) silently. Going forward, meta is the only place for custom user-defined attributes; everything else becomes strongly typed and strictly validated:

models:
  - name: my_model
    description: A model in my project.
    config:
      meta:
        dbt_is_awesome: true
    columns:
      - name: my_column
        config:
          meta:
            some_key: some_value

Duplicate keys in the same YAML file. dbt™ previously used the last occurrence silently. Delete unused duplicates or split them into separate files. Unexpected Jinja blocks. Orphaned blocks such as an {% endmacro %} with no matching {% macro %} were silently ignored; delete them. Properties moving entirely to configs. freshness, meta, tags, docs, group, and access move under config:. For example, source freshness:

sources:
  - name: ecom
    schema: raw
    description: E-commerce data for the Jaffle Shop
    config:
      freshness:
        warn_after:
          count: 24
          period: hour

The --models / --model / -m flag. Renamed to --select / -s back in dbt™ 0.21; it now raises a warning in 1.10 and errors in the v2 engine. Update Bolt schedule commands and scripts that still use it. Custom output path for source freshness. Overriding the sources.json path with --output / -o on dbt source freshness is deprecated; use --target-path to set the artifact path for the whole step. warn_error_options renames. include and exclude are deprecated in favor of error and warn:

flags:
  warn_error_options:
    error: # Previously called "include"
    warn: # Previously called "exclude"
    silence: # To silence or ignore warnings
      - NoNodesForSelectionCriteria

Adapter-specific changes

Snowflake: the platform_detection_timeout_seconds profile parameter controls how long the connector waits when detecting the cloud platform.
BigQuery: jobs that exceed their configured timeout receive a cancellation request; if the request fails, the BigQuery job may keep running in the background until you cancel it manually.

Quick hits

The --use-fast-test-edges flag on dbt build reduces the number of test edges in the execution graph, which can significantly cut run times and memory usage in large projects.
The loaded_at_query source property specifies custom SQL for the source freshness timestamp, as an alternative to loaded_at_field (you cannot set both).
config.meta_get() and config.meta_require() access custom configurations stored under meta.

Upgrade steps

Switch a test environment to 1.10

Pin a Bolt environment to dbt™ 1.10 in Settings > Environments, or switch the version in a staging workspace, and run dbt parse in the Code IDE terminal to surface the new deprecation warnings.

Clean up deprecation warnings

Work through the warnings: move custom properties under config.meta, move freshness, meta, tags, docs, group, and access under config:, relocate standalone anchors, remove duplicate keys and orphaned Jinja blocks, and replace -m with -s in Bolt commands. The dbt-autofix tool or the DinoAI upgrade agent automates most of this.

Guard your warn-error configuration

If any schedule runs with --warn-error or --warn-error-options, decide whether deprecations should fail it. Set "warn": ["Deprecations"] to keep them as warnings while you migrate.

Switch the workspace version

In Settings > Workspace > General, select Change version, choose 1.10, and save. See Upgrade dbt Core™ version.

dbt --version reports 1.10.x, Bolt schedules run green, and dbt parse output shows no deprecation warnings you haven’t consciously deferred. The most likely failure is a --warn-error schedule newly failing on deprecation warnings; fix the code or add "warn": ["Deprecations"] to warn_error_options.

Next steps

Upgrade to dbt™ 1.11

UDFs, and deprecation warnings switched on by default.

Upgrade with the DinoAI agent

Automate the deprecation cleanup with an agent-driven PR.

Manage dbt™ version

The workspace setting and per-environment overrides.

Official v1.10 upgrade guide

The upstream dbt™ documentation this guide is based on.

Foundations

Automate with AI agents

Set up & configure

Orchestrate data pipelines

Migrate to Paradime

Develop in Paradime

Learn dbt™

Upgrade to dbt™ 1.10

What can break

What’s new in dbt™ 1.10

The `--sample` flag

The `anchors:` key for standalone YAML anchors

Parsing `catalogs.yml`

New behavior flags

Deprecation warnings

Adapter-specific changes

Quick hits

Upgrade steps

Next steps

Upgrade to dbt™ 1.11

Upgrade with the DinoAI agent

Manage dbt™ version

Official v1.10 upgrade guide

​What can break

​What’s new in dbt™ 1.10

​The --sample flag

​The anchors: key for standalone YAML anchors

​Parsing catalogs.yml

​New behavior flags

​Deprecation warnings

​Adapter-specific changes

​Quick hits

​Upgrade steps

​Next steps

Upgrade to dbt™ 1.11

Upgrade with the DinoAI agent

Manage dbt™ version

Official v1.10 upgrade guide

What can break

What’s new in dbt™ 1.10

The `--sample` flag

The `anchors:` key for standalone YAML anchors

Parsing `catalogs.yml`

New behavior flags

Deprecation warnings

Adapter-specific changes

Quick hits

Upgrade steps

Next steps