While loading the data from file to hive tables Null values are getting inserted.
sqlCon.sql("create table hive_6(id Int,name String) partitioned by (date String) row format delimited fields terminated by ','");
sqlCon.sql("load data local inpath '/home/cloudera/file.txt' into table hive_6 partition(date='19July')");
sqlCon.sql("select * from hive_6").show()
+----+----+------+
| id|name| date|
+----+----+------+
|null|null|19July|
|null|null|19July|
|null|null|19July|
|null|null|19July|
|null|null|19July|
|null|null|19July|
|null|null|19July|
+----+----+------+
I was facing the same issue when I was reading data from parquet files.
The hive queries will give the correct data, although spark-sql will show null values.
The reason is schema,you should have following-
Firstly-- The column names in the file(txt/parquet) you are reading should be all in lowercase.
Secondly-- The column names in the hive table that you have created should exactly be same as that of the file you are reading.
Thirdly-- The datatypes in both txt/parquet files and hive table should be same.
Related
If I have a csv like -
colName1,colName2
col1Value,col2Value
and a hive ddl like -
CREATE EXTERNAL TABLE tableName (
col2 STRING,
col1 STRING
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
LOCATION 'hdfs://location/to/testcsv/directory'
tblproperties ("skip.header.line.count"="1");
//select col2 from tableName; gives col1Value
This is obviously because in case of text files hive matches column to data field by ordinal position matching. If the underlying file is parquet then the match is done using column names.
I was wondering is there is a hive SerDe someone has written or maybe some SerDe property I am missing that tells hive to map data field names with hive table column names, such that in above example it would return "col2Value" when col2 is queried, even though ordinal position of col2 in hive table and data file does not match.
Thanks in advance!
I am looking to encode columns of a table in hive.
I tried:
hive> create table encode_test(id int, name STRING, phone STRING, address STRING)
> ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
> WITH SERDEPROPERTIES ('column.encode.columns'='phone,address', 'column.encode.classname'='org.apache.hadoop.hive.serde2.Base64WriteOnly') STORED AS TEXTFILE;
Say i have a CSV file, with following row
100,'navis','010-0000-0000','Seoul Seocho'
Now i tried to use.
LOAD DATA LOCAL INPATH
'/home/path/to/csv/test.csv'
INTO TABLE encode_test;
But when doing Select * from encode_test i am getting all columns NULL
Whereas the result should have been
100 navis MDEwLTAwMDAtMDAwMA== U2VvdWwsIFNlb2Nobw==
Also i want to give Fields TERMINATED BY ',' IN create table encode_test query.
but i am getting error: EOF error Near Fields
I also tried creating another table sample
create table sample(id int, name STRING, phone STRING, address STRING)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY ','
LINES TERMINATED BY '\n'
STORED AS TEXTFILE;
And then imported the csv file in the sample table. and it was successfully imported.
then i tried using.
insert into encode_test select * from sample;
But i am getting this new error
Permission denied: user=root, access=WRITE, inode="/user":h dfs:supergroup:drwxr-xr-x
at org.apache.hadoop.hdfs.server.namenode.DefaultAuthorizationProvider.c heckFsPermission(DefaultAuthorizationProvider.java:279)
I'n new into hadoop
Please refer to this link from where i tried this problem
In Hive DDL, ROW FORMAT SERDE and FIELDS TERMINATED BY cannot co-exist together. Instead you can use, field.delim serde property.
create table encode_test(id int, name STRING, phone STRING, address STRING)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
WITH SERDEPROPERTIES (
'field.delim'=',',
'column.encode.columns'='phone,address',
'column.encode.classname'='org.apache.hadoop.hive.serde2.Base64WriteOnly')
STORED AS TEXTFILE;
And for the PermissionDenied exception, run the hive queries as either hdfs or hive user since root user does not have WRITE access to HDFS.
I created below mention external table..
create external table if not exists sensor.building1 (BuildingID int,BuildingMgr string , BuildingAge string, HVACproduct string , Country string) row format delimited fields terminated by ',';
Loaded the table by using below query..
load data inpath '/user/cloudera/sensor/SensorFiles/building.csv' into table sensor.building1;
When I am trying to retrieve the buildingID column using below query, but I am getting null value..
select a.BuildingID
from sensor.building1 as a
limit 10;
Please guide me where I am doing something wrong
You are trying to load a CSV file into hive table but hive's default field delimiter is '\001'
So while you tring to load data from csv (I am assuming its ',' separated) its get failed.
You can create table like :
create external table test1(country string, name string) ROW FORMAT DELIMITED FIELDS TERMINATED BY ',' LINES TERMINATED BY '\n';
I have created a table in Hive with the following query:
create table if not exists employee(CASE_NUMBER String,
CASE_STATUS String,
CASE_RECEIVED_DATE DATE,
DECISION_DATE DATE,
EMPLOYER_NAME STRING,
PREVAILING_WAGE_PER_YEAR BIGINT,
PAID_WAGE_PER_YEAR BIGINT,
order_n int) partitioned by (JOB_TITLE_SUBGROUP STRING) row format delimited fields terminated by ',';
I tried loading data into the create table using below query:
LOAD DATA INPATH '/salary_data.csv' overwrite into table employee partition (JOB_TITLE_SUBGROUP);
For the partitioned table, I have even set following configuration :
set hive.exec.dynamic.partition.mode=nonstrict;
set hive.exec.dynamic.partition=true;
But I am getting below error while executing the load query:
Your query has the following error(s):
Error while compiling statement: FAILED: SemanticException org.apache.hadoop.hive.ql.metadata.HiveException: MetaException(message:Invalid partition key & values; keys [job_title_subgroup, ], values [])
Please help.
If you want to load data into a Hive partition, you have to provide the value of the partition itself in the LOAD DATA query. So in this case, your query would be something like this.
LOAD DATA INPATH '/salary_data.csv' overwrite into table employee partition (JOB_TITLE_SUBGROUP="Value");
Where "Value" is the name of the partition in which you are loading your data. The reason is because Hive will use "Value" to create the directory in which your .csv is going to be stored, which will be something like this: .../employee/JOB_TITLE_SUBGROUP=Value. I hope this helps.
Check the documentation for details on the LOAD DATA syntax.
EDITED
Since the table has dynamic partition, one solution would be loading the .csv into an external table (e.g. employee_external) and then execute an INSERT command like this:
INSERT OVERWRITE INTO TABLE employee PARTITION(JOB_TITLE_SUBGROUP)
SELECT CASE_NUMBER, CASE_STATUS, (...), JOB_TITLE_SUBGROUP
FROM employee_external
I might be little late to reply but can try below steps:
Set below properties first :
Ø set hive.exec.dynamic.partition.mode=nonstrict;
Ø set hive.exec.dynamic.partition=true;
Create temp table first:
CREATE EXTERNAL TABLE IF NOT EXISTS employee_temp(
ID STRING,
Name STRING,
Salary STRING)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY ','
LINES TERMINATED BY '\n'
tblproperties ("skip.header.line.count"="1");
Load Data in temporary table:
hive> LOAD DATA INPATH 'filepath/employee.csv' OVERWRITE INTO TABLE employee;
Create Partitioned Table:
CREATE EXTERNAL TABLE IF NOT EXISTS employee_part(
ID STRING,
Name STRING)
PARTITIONED BY (Salary STRING)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY ','
LINES TERMINATED BY '\n'
tblproperties ("skip.header.line.count"="1");
Load Data into partitioned table from intermediate / temp table:
INSERT OVERWRITE TABLE employee_part PARTITION (SALARY) SELECT * FROM employee;
I have a log file "sample.log" which looks like below:
41 Texas 2000
42 Louisiana4 3211
43 Texas 5000
22 Iowa 4998p
In the log file first column is id, second state name and third amount. If you see State name it has Louisiana4 and sales total it has 4998p. How can I cleanse it so I can insert it into Hive (using Python or other way?). Could you please show the steps?
I want to insert into Hive table tblSample:
Table schema is:
CREATE TABLE tblSample(
id int,
state string,
sales int)
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t'
STORED AS TEXTFILE
LOCATION '/user/cloudera/Staging'
;
To load data into Hive table I could do:
load data local inpath '/home/cloudera/sample.log' into table tblSample;
Thank you!
You could load data as is into a hive table and then use UDFs to cleanse data and load into another table. This would be far more efficient than Python as it will be running as a mapr reduce.
I would rather store the data as it is and do the cleansing while fetching the data. It would be much simpler. No external code required. For example :
hive> CREATE TABLE tblSample(
> id string,
> state string,
> sales string)
> ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t'
> STORED AS TEXTFILE
> LOCATION '/user/cloudera/Staging';
hive> select regexp_replace(state, "[0-9]", ""), regexp_replace(sales, "[a-z]", "") from tblSample;
HTH