Databricks: SQL Pivot does not work - but Python does

I am trying to pivot a SQL table in Databricks on the FeatureName column using the FeatureDescription column as the value.

But somehow the SQL query that I have resolves to null values being populated across the table.

Here is the original table:

Here is the SQL query:

SELECT *
FROM
  (SELECT AddressKey, FeatureName, FeatureDescription
   FROM radp_dev_cat.rdp_svr_sales.vw_customerfeatures) src
PIVOT (
  max(FeatureDescription)
  FOR FeatureName IN (
    'Carcasspricecategory' AS Carcasspricecategory,
    'Channel' AS Channel,
    'Class' AS Class,
    'Contractor' AS Contractor,
    'DebtorCategory' AS DebtorCategory,
    'DomesticCustTradingasExport' AS DomesticCustTradingasExport,
    'FarmingOperations' AS FarmingOperations,
    'FarmingProductionType' AS FarmingProductionType,
    'FeatureName' AS FeatureName,
    'FMInternalRegions' AS FMInternalRegions,
    'Groups' AS Groups,
    'MillITSupplierType' AS MillITSupplierType,
    'NationalGroups' AS NationalGroups,
    'PaymentMethod' AS PaymentMethod,
    'PurchaseGroup' AS PurchaseGroup,
    'RegisteredforGST' AS RegisteredforGST,
    'SalesGroup' AS SalesGroup,
    'SiteBaxters' AS SiteBaxters,
    'SiteCorowaMeats' AS SiteCorowaMeats,
    'SiteDVP' AS SiteDVP,
    'SiteFeedmill' AS SiteFeedmill,
    'SiteNSWWholesale' AS SiteNSWWholesale,
    'Type' AS Type
  )
);

The result:

All null values for all columns across the table.

I tried using any, first, max, etc. for the aggregate function. Same result.

But it works why I do the same thing using Python:

%python
import pandas as pd

df = spark.sql('SELECT * FROM  radp_dev_cat.rdp_svr_sales.vw_customerfeatures')

df_pivot = df.groupBy('AddressKey').pivot('FeatureName').agg({"FeatureDescription": "first"}).toPandas()

df_pivot

Why does this happen? Is something wrong with my SQL query?

Thanks.

The problem was the underlying view that had the FeatureName, FeatureDescription data. The columns had trailing whitespace in the records so none of the PIVOT columns matched.

I trimmed the underlying view and it worked.

Thanks for the insight @nbk.