排查Hive报错:org.apache.hadoop.hive.serde2.SerDeException: java.io.IOException: Start of Array expected

CREATE TABLE json_nested_test (
    count string,
    usage string,
    pkg map<string,string>,
    languages array<string>,
    store map<string,array<map<string,string>>>)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;

以上述sql创建表json_nested_test后,查询时发现报错:Failed with exception java.io.IOException:org.apache.hadoop.hive.serde2.SerDeException: java.io.IOException: Start of Array expected

于是进行如下测试,发现是org.apache.hive.hcatalog.data.JsonSerDe 对复杂类型支持不足造成,例如map<string,array<string>>,这个例子中就是不支持array作为map的value.
CREATE TABLE json_nested_test_openx (
    count string,
    usage string,
    pkg map<string,string>,
    languages array<string>,
    store map<string,array<map<string,string>>>)
ROW FORMAT SERDE 'org.openx.data.jsonserde.JsonSerDe'
STORED AS TEXTFILE;

CREATE TABLE s1 (
    count string,
    usage string
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s1.txt' overwrite into table s1;
select * from s1;
CREATE TABLE s2 (
    count string,
    usage string,
    pkg map<string,string>
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s2.txt' overwrite into table s2;
select * from s2;

CREATE TABLE s3 (
    count string,
    usage string,
    pkg map<string,string>,
    languages array<string>
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s3.txt' overwrite into table s3;
select * from s3;
CREATE TABLE s4 (
    count string,
    usage string,
    pkg map<string,string>,
    languages array<string>,
    store map<string,array<map<string,string>>>
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s4.txt' overwrite into table s4;
select * from s4;

CREATE TABLE s5 (
    store map<string,array<map<string,string>>>
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s5.txt' overwrite into table s5;
select * from s5;

CREATE TABLE s6 (
    store map<string,array<string>>
)
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
STORED AS TEXTFILE;
load data local inpath '/home/work/s6.txt' overwrite into table s6;
select * from s6;

这个Serde hive自带,路径为$HIVE_HOME/hcatalog/share/hcatalog/hive-hcatalog-core-1.2.1.jar,它也存在其它问题:不支持数据文件中的空行

原文地址:https://www.cnblogs.com/aprilrain/p/6916359.html