Requirement
You have a file which is delimited by multiple characters (%$) and you want to create a table in the hive on top of it.
Solution
Step 1: Sample File
Create a sample file named as sample_1.txt.
Download from here sample_1
(You can skip this step if you already have a file, just place it into a local directory.)
If you have any sample data with you, then put the content in that file with a delimiter (%$). If you have created a file in windows, then transfer it to your Linux machine via WinSCP.
Please refer below screenshot.
I have a local directory named as /root/local_bdp/posts/load-multi-character-delimited-file-into-hive, so I have placed a sample_1.txt file in that directory.
Step 2: Copy File to HDFS
Run Below commands in the shell for initial setup.
First, create a Hdfs directory hv_mlt_chr using below command:
hadoop fs -mkdir hv_mlt_chr
Put the file in Hdfs using below command:
hadoop fs -put /root/local_bdp/posts/load-multi-character-delimited-file-into-hive/sample_1.txt hv_mlt_chr/
Check whether the file is available in Hdfs or not using below command:
hadoop fs -ls hv_mlt_chr/
NOTE: – For me, the default Hdfs directory is /user/root/
You can check the content of the file using below command:
hadoop fs -cat hv_mlt_chr/sample_1.txt
Step 3: Create Hive Table and Load data
Now, you have a file in Hdfs, you just need to create an external table on top of it.
Use below hive script to create an external table userTrck in schema bdp. Run below script in hive CLI.
CREATE SCHEMA IF NOT EXISTS bdp;
CREATE EXTERNAL TABLE bdp.userTrck
(id STRING,
u_type STRING,
s_code STRING)
ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.MultiDelimitSerDe'
WITH SERDEPROPERTIES ("field.delim"="%$")
LOCATION 'hdfs://sandbox.hortonworks.com:8020/user/root/hv_mlt_chr';
You can change the delimiter as per your requirement. In this case, delimiter should be %$. Make sure that you have written serde2.MultiDelimitSerDe.
Step 4: Verify data
Please check whether data is showing in a table or not using below command:
select * from bdp.userTrck;
Wrapping Up
Sometimes data comes from a system which gives multi-character delimited file. It may come from website or any other logging system, generally we use multi-character as delimiter when data already have (or may have) all possible special symbol, that time separating one column with other would be tough, in that case we can use multi characters as delimiter, which is impossible to occur in the data as value somewhere in the string.
Keep Learning.
Don’t miss the tutorial on Top Big data courses on Udemy you should Buy