pyspark操作hive分区表及.gz.parquet和part-00000文件压缩问题

df.write.saveAsTable("表名",mode='overwrite')

# append 写入
df.repartition(1).write.partitionBy('dt').insertInto("表名")
# overwrite 写入
df.repartition(1).write.partitionBy('dt').insertInto("表名",overwrite=True)
# 动态分区使用该方法

-- 使用snappy
CREATE TABLE if not exists ods.table_test(
    id string,
    open_time string
	)
COMMENT '测试'
PARTITIONED BY (`dt` string COMMENT '按天分区')
row format delimited fields terminated by '\001' 
STORED AS PARQUET 
TBLPROPERTIES ('parquet.compression'='SNAPPY');

-- 使用gzip
CREATE TABLE if not exists ods.table_test(
    id string,
    open_time string
	)
COMMENT '测试'
PARTITIONED BY (`dt` string COMMENT '按天分区')
row format delimited fields terminated by '\001' 
STORED AS PARQUET 
TBLPROPERTIES ('parquet.compression'='GZIP');
 
-- 使用uncompressed
CREATE TABLE if not exists ods.table_test(
    id string,
    open_time string
	)
COMMENT '测试'
PARTITIONED BY (`dt` string COMMENT '按天分区')
row format delimited fields terminated by '\001' 
STORED AS PARQUET 
TBLPROPERTIES ('parquet.compression'='UNCOMPRESSED');

 
-- 使用默认
CREATE TABLE if not exists ods.table_test(
    id string,
    open_time string
	)
COMMENT '测试'
PARTITIONED BY (`dt` string COMMENT '按天分区')
row format delimited fields terminated by '\001' 
STORED AS PARQUET;
 
-- 设置参数 set parquet.compression=SNAPPY;

drop table if exists ods.table_test
CREATE TABLE if not exists ods.table_test(
id string,
open_time string
)
COMMENT '测试'
PARTITIONED BY (`dt` string COMMENT '按天分区')
row format delimited fields terminated by '\001' 
STORED AS PARQUET 
TBLPROPERTIES ('parquet.compression'='GZIP');

rdd.saveAsTextFile('hdfs://表全路径')