velox icon indicating copy to clipboard operation
velox copied to clipboard

Analyze Hive Parquet table failed

Open xumingming opened this issue 1 year ago • 3 comments

Bug description

presto:tpch100g> analyze orders;

Query 20240121_133327_00322_v5a3c, FAILED, 3 nodes
Splits: 69 total, 0 done (0.00%)
[Latency: client-side: 93ms, server-side: 76ms] [0 rows, 0B] [0 rows/s, 0B/s]

Query 20240121_133327_00322_v5a3c failed: scalar type has no children

System information

N/A

Relevant logs

No response

xumingming avatar Jan 21 '24 13:01 xumingming

@xumingming Will you be able to paste the stack please?

yingsu00 avatar Feb 02 '24 17:02 yingsu00

@yingsu00 The whole stack trace(from Presto UI) is:

std::exception: scalar type has no children

No log in the native worker's log.

More clue:

  • analyze customer success.
  • analyze lineitem fail.

xumingming avatar Feb 03 '24 02:02 xumingming

I have same problem. analyse table on presto native execution ;

yixi-gu avatar Apr 26 '24 07:04 yixi-gu

I ran analyze locally on a hive parquet table and it ran successfully:

presto:tpcds> analyze catalog_sales;
ANALYZE: 89807 rows

Query 20240516_053830_00003_ecv5c, FINISHED, 2 nodes
Splits: 26 total, 26 done (100.00%)
[Latency: client-side: 0:01, server-side: 0:01] [89.8K rows, 9.56MB] [65K rows/s, 6.92MB/s]

presto:tpcds> show create table catalog_sales;
                Create Table                 
---------------------------------------------
 CREATE TABLE hive.tpcds.catalog_sales (     
    "cs_sold_date_sk" bigint,                
    "cs_sold_time_sk" bigint,                
    ..
    "cs_coupon_amt" decimal(7,2),            
    "cs_ext_ship_cost" decimal(7,2),         
    "cs_net_paid" decimal(7,2),              
    "cs_net_paid_inc_tax" decimal(7,2),      
    "cs_net_paid_inc_ship" decimal(7,2),     
    "cs_net_paid_inc_ship_tax" decimal(7,2), 
    "cs_net_profit" decimal(7,2)             
 )                                           
 WITH (                                      
    format = 'PARQUET'                       
 )                                           
(1 row)

I can reproduce this issue if use-alternative-function-signatures=false. Please set this flag to true in config.properties.

presto:tpcds> analyze catalog_sales;

Query 20240516_054226_00000_xwpwz, FAILED, 2 nodes
Splits: 9 total, 0 done (0.00%)
[Latency: client-side: 0:04, server-side: 0:03] [0 rows, 0B] [0 rows/s, 0B/s]

Query 20240516_054226_00000_xwpwz failed: scalar type has no children

karteekmurthys avatar May 16 '24 05:05 karteekmurthys