infromation_schema.columns not populated correctly

[guillesd]

When creating a table and then querying information_schema.columns, the column_name and the data_type fields are not populated for Iceberg tables.

-- Attach catalog (following setup of this repo)
CREATE SECRET (
      TYPE S3,
      KEY_ID 'admin',
      SECRET 'password',
      ENDPOINT '127.0.0.1:9000',
      URL_STYLE 'path',
      USE_SSL 0
  );
ATTACH '' AS my_datalake (
      TYPE ICEBERG,
      CLIENT_ID 'admin',
      CLIENT_SECRET 'password',
      ENDPOINT 'http://127.0.0.1:8181'
  );
CREATE SCHEMA my_datalake.default;
USE my_datalake.default;
CREATE TABLE t AS SELECT range a FROM range (4);
SELECT table_catalog, table_schema, table_name, column_name, data_type FROM information_schema.columns WHERE table_catalog='my_datalake' AND table_name='t';
┌───────────────┬──────────────┬────────────┬─────────────┬───────────┐
│ table_catalog │ table_schema │ table_name │ column_name │ data_type │
│    varchar    │   varchar    │  varchar   │   varchar   │  varchar  │
├───────────────┼──────────────┼────────────┼─────────────┼───────────┤
│ my_datalake   │ default      │ t          │ __          │ UNKNOWN   │
└───────────────┴──────────────┴────────────┴─────────────┴───────────┘

[Tmonster]

This is by design. If you have N tables, to get the column information for every table you have to make N requests. So 100 tables means 100 requests to the Load Table Information endpoint, which will slow down table listing dramatically and make DuckDB-Iceberg almost unusable. PR here

If you do want column information, Describe {table_name} will show you all column information.