外部非分区 Hive 表显示 0 条记录，即使 parquet 文件有数据，当使用 spark.read.parquet 读取时，它显示正确的数据

Question

I have an external non-partitioned hive table created on Azure DataBricks.

CREATE EXTERNAL TABLE reference.date_file(
  `date_init` string,
  `trade_date_minus_5` string,
  `trade_date_minus_4` string,
  `trade_date_minus_3` string,
  `trade_date_minus_2` string,
  `trade_date_minus_1` string,
  `trade_date` string,
  `trade_date_plus_1` string,
  `trade_date_plus_2` string,
  `trade_date_plus_3` string,
  `trade_date_plus_4` string,
  `trade_date_plus_5` string)
STORED AS PARQUET
LOCATION '/mnt/data/reference/data/DATE_FILE'
TBLPROPERTIES (
  'parquet.compress'='SNAPPY'
)

It also has parquet file present in the location which I copied from other location.

When I do select * from reference.date_file it returns 0 results even if data/records are present.

The same file when I do...

val temp = spark.read.parquet("/mnt/data/reference/data/DATE_FILE")
temp.show()

... returns the records.

Can someone please help what can be the issue here?

Answer 1

This is solved.

I simply changed the location to some other location, re-created the table, restored the original location, recreated the table. And it started showing the data. May be there was some issue with the hive metadata. Changing the location and restoring corrected the metadata I guess.