Yes, you can exclude a specific column (such as

internal_id

) from the

elementary.all_columns_anomalies

test by using the

exclude_prefix

or

exclude_regexp

parameters. For a single column, the simplest way is to use

exclude_prefix

if your column name has a unique prefix, or

exclude_regexp

for an exact or partial match. For example, to exclude only the

internal_id

column, you can use:

yaml models:   - name: your_model_name     tests:       - elementary.all_columns_anomalies:           exclude_regexp: "^internal_id$"

This configuration will ensure that only the

internal_id

column is excluded from the analysis, while all other columns are still monitored. The

exclude_regexp

parameter matches the column name exactly in this case. Alternatively, if you want to exclude columns by prefix (e.g., all columns starting with

internal_

), you could use:

yaml       - elementary.all_columns_anomalies:           exclude_prefix: "internal_"

But for a single column, the regex approach is more precise. Note that

exclude_prefix

and

exclude_regexp

are only available for the

all_columns_anomalies

test, not for

volume_anomalies

or other tests [exclude_regexp documentation](https://docs.elementary-data.com/data-tests/anomaly-detection-configuration/exclude_regexp) [exclude_prefix documentation](https://docs.elementary-data.com/data-tests/anomaly-detection-configuration/exclude_prefix). If you want to set the

seasonality

parameter (e.g., from day to week), you can add it as follows:

yaml       - elementary.all_columns_anomalies:           exclude_regexp: "^internal_id$"           seasonality: day_of_week

Let me know if you need an example for another test or further clarification!

Based on the knowledge sources, your configuration for excluding the

internal_id

column using

exclude_regexp: "^internal_id$"

in the

elementary.all_columns_anomalies

test is correct and should work for excluding a single column by exact name match [exclude_regexp documentation](https://docs.elementary-data.com/data-tests/anomaly-detection-configuration/exclude_regexp). However, there are a few important caveats and known issues: 1. YAML Indentation and Syntax: Ensure your YAML indentation is correct. The

exclude_regexp

parameter should be at the same level as other test parameters. 2. Test Name Still References Excluded Column: Even when a column is excluded, the generated dbt test name may still include the column name due to how dbt constructs test names. This does not necessarily mean the column is being tested, but it can be confusing [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1650985870725969). 3. dbt Package Version: The ability to exclude columns using

exclude_regexp

was added in version 0.3.14. If you are using an older version, this feature will not work. Please ensure you are using at least version 0.3.14 [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1650985870725969). 4. Known Limitation – Only One Prefix/Regexp: Only one

exclude_prefix

or

exclude_regexp

can be used at a time. If you need to exclude multiple columns, you must use a regular expression that matches all of them, e.g.,

exclude_regexp: "^(internal_id|other_column)$"

 [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1650985870725969). 5. Test Still Fails on Excluded Column: If the test is still failing on

internal_id

, it could be due to: - A typo in the column name. - The column is not being matched by the regex (double-check the exact column name in your database). - The dbt package version is too old. - There is a bug or edge case in the implementation (as reported by other users) [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1697205851749239). What you can try: - Double-check the spelling and case of

internal_id

in your model. - Make sure you are using dbt-elementary version 0.3.14 or newer. - Try a broader regex, e.g.,

exclude_regexp: "internal_id"

(without the anchors) to see if it matches. - If you are still seeing the error, it may be a bug. Other users have reported similar issues, and the workaround is to use

exclude_regexp

with a pattern that matches the column, or to define the test at the column level and exclude the problematic column [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1697205851749239). If none of these resolve the issue, it is likely a bug, and you may want to open a ticket on the Elementary GitHub repository as other users have done. Let me know if you need help with a specific regex pattern or further troubleshooting!

Your YAML format is almost correct, but there are indentation issues that could cause problems with dbt and Elementary. The parameters under each test should be indented properly. Here’s the corrected format based on the knowledge sources:

yaml models:   - name: test_upload_on_day     description: Middle layer model     config:       elementary:         timestamp_column: report_day     tests:       - elementary.volume_anomalies:           tags: ["elementary"]       - elementary.all_columns_anomalies:           tags: ["elementary"]           exclude_regexp: "^internal_id$"           seasonality: day_of_week

Key points: - The parameters (

tags

,

exclude_regexp

,

seasonality

) should be indented two spaces under the test name. - Your use of

exclude_regexp: "^internal_id$"

is correct for excluding a single column by exact name [exclude_regexp documentation](https://docs.elementary-data.com/data-tests/anomaly-detection-configuration/exclude_regexp). - The

seasonality

parameter is also correctly placed. Version compatibility: You mentioned you are using Elementary version 0.19.5. The knowledge sources confirm that

exclude_regexp

is supported in recent versions, but the feature was introduced in version 0.3.14 of the dbt package. If you are referring to the dbt package version, ensure it is at least 0.3.14 for this feature to work as expected [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1650985870725969). If you still encounter issues after correcting the indentation and confirming the version, it may be a bug or an edge case as reported by other users. Let me know if you need further help!

Yes, your understanding is correct and aligns with the guidance from the knowledge sources. If you want to exclude specific columns (like

internal_id

) from anomaly detection, and the

exclude_regexp

or

exclude_prefix

options are not working as expected or are too broad, you can manually split the test by applying

elementary.column_anomalies

only to the columns you want to monitor. This approach gives you full control over which columns are tested. Here’s how you can do it:

yaml models:   - name: test_upload_on_day     config:       elementary:         timestamp_column: report_day     columns:       - name: column1         tests:           - elementary.column_anomalies:               # your desired monitors here       - name: column2         tests:           - elementary.column_anomalies:               # your desired monitors here       # Do NOT include internal_id here

This method is explicitly mentioned as a workaround in the knowledge sources, especially when you want to avoid running tests on certain columns and the built-in exclusion parameters are insufficient or problematic [Slack discussion](https://slack.com/archives/C02CTC89LAX/p1650985870725969). So, replacing

all_columns_anomalies

with multiple

column_anomalies

tests for only the desired columns (excluding

internal_id

) is a valid and supported approach.