# Adding new columns to a snapshot

**URL:** <https://discourse.getdbt.com/t/adding-new-columns-to-a-snapshot/18092>\
**Category:** Help\
**Created:** [February 7, 2025, 7:24pm UTC](https://discourse.getdbt.com/t/adding-new-columns-to-a-snapshot/18092 "2025-02-07T19:24:24Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![chadwicke619](https://avatars.discourse-cdn.com/v4/letter/c/839c29/32.png) [@chadwicke619](https://discourse.getdbt.com/u/chadwicke619)\
**Post date:** [February 7, 2025, 7:24pm UTC](https://discourse.getdbt.com/t/adding-new-columns-to-a-snapshot/18092/1 "2025-02-07T19:24:24Z")

</div>

I have the following snapshot:

```auto
{% snapshot meter_readings_snapshot %}

    {{
        config(
          target_schema=get_snapshot_schema(),
          unique_key='id',

          strategy='timestamp_with_nulls',
          updated_at='updated_at',
          invalidate_hard_deletes=True,
        )
    }}

select id,
       meter_id,
       estar_meter_reading_id,
       from_date,
       to_date,
       cost,
       usage,
       exported,
       demand,
       demand_charge,
       is_custom,
       created_at,
       is_estimate,
       updated_at,
       _sdc_extracted_at,
       _sdc_batched_at,
       _sdc_deleted_at

from {{ source('CORE', 'METER_READINGS') }}

{% endsnapshot %}

```

We added two new columns to this METER\_READINGS table in Production (updated\_by and created\_by), but now I would like to add those columns to the snapshot. I expected that the next time dbt ran it would pull in those new columns, pull in the _current_ values for the columns, and then all the historical data where dbt\_valid\_to is not null would have null values for the new fields. That’s not what happened. Instead I got two new columns that were almost entirely null, and then I’m slowly getting values, presumably as people make changes, again presumably because of the update strategy, I imagine?

So… my question is, how do I need to go about this so that I preserve all the historical data we’ve been collecting for years, while also bringing in the current values of the two new columns for all the current rows? Thanks all.