在 CREATE TABLE 的輸入中沒有可行的替代方法，列定義為 MAP <int,<array<map<int,float> &gt;&gt;&gt; </int,<array<map<int,float>

Question

我正在運行以下火花查詢

DROP TABLE IF EXISTS hmi_feature_array;

CREATE TABLE IF NOT EXISTS hmi_feature_array (
    cycle_timestamp BIGINT,
    campaign_id BIGINT,
    arm ARRAY<BIGINT>,
    context_feature MAP<INT,<ARRAY<MAP<INT,FLOAT>>>>
)
PARTITIONED BY(ds STRING)
TBLPROPERTIES (
    'RETENTION' = '90',
    'noUII' = '1'
);

但是，我在context_feature MAP<INT,<ARRAY<MAP<INT,FLOAT>>>> no viable alternative at input 。 完整的相關錯誤日志如下：

[INFO ][1479233][2022-08-23 15:44:33,685][MainThread][utils.py:514 ]- Logging Spark App Metrics to XDB through wrapper: Sample(<class 'dict'>, {'normal': {'exception': "org.apache.spark.sql.catalyst.parser.ParseException: \nno viable alternative at input 'CREATE TABLE IF NOT EXISTS hmi_feature_array (\\n cycle_timestamp BIGINT,\\n campaign_id BIGINT,\\n arm ARRAY<BIGINT>,\\n context_feature MAP<INT,<'(line 5, pos 28)\n\n== SQL ==\nCREATE TABLE IF NOT EXISTS hmi_feature_array (\n cycle_timestamp BIGINT,\n campaign_id BIGINT,\n arm ARRAY<BIGINT>,\n context_feature MAP<INT,<ARRAY<MAP<INT,FLOAT>>>>\n----------------------------^^^\n)\nPARTITIONED BY(ds STRING)\nTBLPROPERTIES

為什么會這樣？

Answer 1

事實證明，它應該是context_feature MAP<INT,ARRAY<MAP<INT,FLOAT>>> <INT,FLOAT>>>> 而不是context_feature MAP<INT,<ARRAY<MAP<INT,FLOAT>>>> 。 基本上ARRAY周圍不需要<> 。