hive-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Grover <>
Subject Re: Hive - Load file with delimiters
Date Wed, 28 Nov 2012 16:17:08 GMT
Moving to since this is an Apache Hive related question

Hi Ram,
Hive doesn't provide a way to do this out of the box. I have seen other
people have the same request before (and more, e.g. use of escape
characters to embed the enclosing characters, etc.) and have just created a
Hive JIRA to discuss and implement this.

However, there are several ways I can think of to work around this for now.
They are:
1. Pre-process the files to get rid of the enclosing characters ("). This
may or may not be scalable depending on your implementation, would be
fairly simple to implement.
2. Create a staging Hive table with enclosing characters as a part of the
data. Then, you could do one of 3 things with this staging table.
 2a. Create another Hive table that gets populated by a Hive query on the
staging table. The query basically strips out the enclosing characters.
 2b. Create a view on the existing Hive staging table that presents the
data after removing the enclosing characters.
 2c. Use the existing RegexSerDe when creating the Hive staging table to
get rid of the enclosing characters when serializing/deserializing the data
from the staging table.
3. Create a custom informat or SerDe that supports enclosing characters in
delimited fields.

#3 above will be an answer to the JIRA I just created. So, the community
would be grateful to you if you go that route and contribute it back. Of
course, feel free to choose another option if that works better for you.

For your reference, there is a similar thread that I had replied to at:

Good luck!

On Wed, Nov 28, 2012 at 7:25 AM, Ram Krishnamurthy <> wrote:

> I am having issues loading a delimited csv. here is a sample data
> "1", "123456", " ", "Charlotte, NC"
> The delimiter is ,
> and fields are enclosed by " "
> Does anyone have syntax for loading this in Hive?
> Thanks, Ram

View raw message