june_to_may_id_mapping missing base/land_cover?

[mikemoraned]

Hi there, I have some ids in my config and I just was updating my project to use the new ids in the June release.

One of my ids was 08b2db2b70d8bfff0005d68bbb7ae2f8 (Jerusalem area in base/land_cover of 2025-05-21.0 release) and this seems to correspond to e77650e6-4bc3-5e54-b5b4-46f8e3b1b375 (in 2025-06-25.0 release).

However, I had to manually do this mapping using the previous bounding box as a guide, as the june_to_may_id_mapping S3 (s3://overturemaps-extras-us-west-2/june_to_may_id_mapping/) only contains data for four top-level themes:

aws s3 ls --region us-west-2 --no-sign-request s3://overturemaps-extras-us-west-2/june_to_may_id_mapping/                
                           PRE theme=buildings/
                           PRE theme=divisions/
                           PRE theme=places/
                           PRE theme=transportation/
2025-06-18 20:20:18          0

Am I doing something wrong in accessing this or are there really only mappings for these themes?

Note that I am not stuck as most of my ids were found in divisions/division_area, but I thought I should mention it.

Thanks,

Mike

[jenningsanderson]

Hi @mikemoraned, Yes, there is no ID mapping file for base features because the majority of these features come directly from OpenStreetMap.

Instead, you can rely on the record_id value in the sources attribute to determine if the feature is the same between releases. Here's the pseudo-code query to construct your own mapping file between releases:

SELECT 
  may.id AS may_id, 
  june.id AS june_id
FROM 
  may JOIN june ON 
SPLIT(may.sources[1].record_id, '@')[1] = SPLIT(may.sources[1].record_id, '@')[1]
  AND may.class = june.class 
  AND may.subtype = june.subtype 
  AND may.type = june.type
WHERE type IN ('infrastructure','water','land','land_use')

Note that we only consider the record_id value before the @ symbol because the feature might have incremented version in OSM.

[danabauer]

Thanks for this excellent question, @mikemoraned. I transferred this issue to our community discussion board because others have asked similar questions and @jenningsanderson provided a very helpful answer.

[mikemoraned]

Hi there! Thanks for the info.

I had a go at what you suggested, but this land_cover area appears to come from "ESA WorldCover" dataset, and the record_id is NULL:

-- May                 
SELECT sources[1].record_id, sources, bbox
FROM base_land_cover_may
WHERE id='08b2db2b70d8bfff0005d68bbb7ae2f8'

-- |(sources[1]).record_id|sources                                                     |bbox                                        |
-- |----------------------|------------------------------------------------------------|--------------------------------------------|
-- |                      |{'[, ESA WorldCover, null, 2024-11-07T00:00:00.000Z, null]'}|[35.156044, 35.507954, 31.656881, 31.952284]|

-- June
SELECT sources[1].record_id, sources, bbox
FROM base_land_cover_june
WHERE id='e77650e6-4bc3-5e54-b5b4-46f8e3b1b375'

-- |(sources[1]).record_id|sources                                                           |bbox                                        |
-- |----------------------|------------------------------------------------------------------|--------------------------------------------|
-- |                      |{'[, ESA WorldCover, null, 2024-11-07T00:00:00.000Z, null, null]'}|[35.156044, 35.507954, 31.656881, 31.952284]|

-- definitions of `base_land_cover_may` and `base_land_cover_june` from my local copies of the OvertureMaps data
CREATE OR REPLACE VIEW base_land_cover_may AS
SELECT
    *
FROM
    read_parquet('/Volumes/PRO-G40/OvertureMaps/data/release/2025-05-21.0/theme=base/type=land_cover/*', 
                 hive_partitioning=1)
                 
CREATE OR REPLACE VIEW base_land_cover_june AS
SELECT
    *
FROM
    read_parquet('/Volumes/PRO-G40/OvertureMaps/data/release/2025-06-25.0/theme=base/type=land_cover/*', 
                 hive_partitioning=1) 

[mikemoraned]

Also, I am confused by this:

you can rely on the record_id value in the sources attribute to determine if the feature is the same between releases

If the geometry refers to the same real feature wouldn't the GERs id always be the same?

(Btw, just in case we are talking at cross-purposes, I should be clearer: I was doing a mapping of changes coming from the one-off breaking change to GERS (https://docs.overturemaps.org/blog/2025/06/25/release-notes/#whats-new-big-changes-to-gers))