Synapse Studio throughput issue

bug_report.sql

-- Querying Delta Lake, 643715 rows
-- Created, optimized and vacuumed with Databricks
-- Partitioned by year and month (filepath 1 and 2)
-- Spark parquet repartition(1) means 1 file per partition
-- Runtime:
--   SSMS: 01:31
--   Synapse Workspace top 5000 rows: 03:15
--   Synapse Workspace all rows: 25:29
-- Hard-coding the partition names makes no discernible difference

SELECT g_l_account_no_, posting_date document_type, amount, R.filepath(1), R.filepath(2), R.filepath(3)
FROM OPENROWSET(
    BULK 'lri-data-lake/data/delta/g_l_entry.parquet/posting_year=*/posting_month=*/*.parquet',
    DATA_SOURCE = 'ADLSStorage',
    FORMAT = 'PARQUET'
)
WITH (
    [timestamp] BIGINT,
    [g_l_account_no_] INT,
    [posting_date] DATETIME2,
    [document_type] INT,
    [document_no_] VARCHAR(40),
    [description] VARCHAR(100),
    [bal_account_no_] VARCHAR(40),
    [amount] DECIMAL(38,20),
    [global_dimension_1_code] VARCHAR(40),
    [global_dimension_2_code] VARCHAR(40),
    [user_id] VARCHAR(40),
    [source_code] VARCHAR(20),
    [system-created_entry] INT,
    [prior-year_entry] INT,
    [job_no_] VARCHAR(40),
    [quantity] DECIMAL(38,20),
    [vat_amount] DECIMAL(38,20),
    [business_unit_code] VARCHAR(20),
    [journal_batch_name] VARCHAR(20),
    [reason_code] VARCHAR(20),
    [gen_posting_type] INT,
    [gen_bus_posting_group] VARCHAR(20),
    [gen_prod_posting_group] VARCHAR(20),
    [bal_account_type] INT,
    [transaction_no_] INT,
    [debit_amount] DECIMAL(38,20),
    [credit_amount] DECIMAL(38,20),
    [document_date] DATETIME2,
    [external_document_no_] VARCHAR(40),
    [source_type] INT,
    [source_no_] VARCHAR(40),
    [no_series] VARCHAR(20),
    [tax_area_code] VARCHAR(40),
    [tax_liable] INT,
    [tax_group_code] VARCHAR(20),
    [use_tax] INT,
    [vat_bus_posting_group] VARCHAR(20),
    [vat_prod_posting_group] VARCHAR(20),
    [additional-currency_amount] DECIMAL(38,20),
    [add_-currency_debit_amount] DECIMAL(38,20),
    [add_-currency_credit_amount] DECIMAL(38,20),
    [close_income_statement_dim_id] INT,
    [ic_partner_code] VARCHAR(40),
    [reversed] INT,
    [reversed_by_entry_no_] INT,
    [reversed_entry_no_] INT,
    [prod_order_no_] VARCHAR(40),
    [fa_entry_type] INT,
    [fa_entry_no_] INT,
    [value_entry_no_] INT,
    [gd3code] VARCHAR(40),
    [gd4code] VARCHAR(40),
    [gd5code] VARCHAR(40),
    [gd6code] VARCHAR(40),
    [gd7code] VARCHAR(40),
    [gd8code] VARCHAR(40),
    [gd9code] VARCHAR(40),
    [gd10code] VARCHAR(40),
    [gd11code] VARCHAR(40),
    [gd12code] VARCHAR(40),
    [gd13code] VARCHAR(40),
    [gd14code] VARCHAR(40),
    [gd15code] VARCHAR(40),
    [gd16code] VARCHAR(40),
    [gd17code] VARCHAR(40),
    [gd18code] VARCHAR(40),
    [gd19code] VARCHAR(40),
    [gd20code] VARCHAR(40),
    [licensee_type] INT,
    [licensee_no_] VARCHAR(40),
    [entry_no_] INT,
    [reconciled] INT,
    [reconciled_per] DATETIME2,
    [reconciled_id] INT,
    [royalty_posting_type_code] VARCHAR(20),
    [ic_partn_irrelevant_for_cons_] INT,
    [rights_sales_no_] VARCHAR(40),
    [rights_owner_no_] VARCHAR(40),
    [rights_buyer_no_] VARCHAR(40),
    [creation_date] DATETIME2,
    [reposting] INT
) [R]
WHERE R.filepath(1) = 2020
AND R.filepath(2) = 7