Currently what is a legal column or table name is a very limited set of ascii characters. This is because hive needs to pass this meta information between components, sometimes as XML sometimes as json, and these serializations are not able to deal with complex encoding. There is some flexibility with comments but different characters sets in the columns and table names is a large change.
On Thu, Aug 23, 2012 at 11:01 AM, Namas Amitabha <hyacinth...@hotmail.com> wrote: > Hi Rekha, > > > > I’ve set the JDBC ConnectionURL like this > “?useUnicode=true&characterEncoding=UTF-8”,but it doesn’t work. > > > > And if my table content contains Chinese words, hive can display normally > when I exec select query, only table comments and column comments are > garbled when I tried to exec “desc table”. > > So I don’t know where is the problem. I even tried to modify some hive > source code in org.apache.hadoop.hive.jdbc to convert comment string to > “UTF-8” charset, but it still doesn’t work. > > > > But thanks all the same. > > > > > > > > Thanks & Regards > > Zhan > > 发件人: Joshi, Rekha [mailto:rekha_jo...@intuit.com] > 发送时间: 2012年8月23日 19:42 > 收件人: user@hive.apache.org; hyacinth...@hotmail.com > 主题: Re: hive comment in Chinese > > > > Hi Zhan, > > > > Ideally there must be some 'hive.cli.print.charset' property in config xml > files somewhere which gets picked up by CliDriver (in say > HiveConf.ConfVars.HIVE_CLI_PRINT_CHARSET), but I don't see it. > > > > There is one in hive-site.xml to specify character set, but I am not sure if > it is picked by CliDriver.But try? > > <property> > > <name>javax.jdo.option.ConnectionURL</name> > > > <value>jdbc:mysql://<your-db>:3306/hive?characterEncoding=UTF-8</value> > > </property> > > > > Thanks > > Rekha > > > > From: NamasAmitabha <hyacinth...@hotmail.com> > Reply-To: <user@hive.apache.org> > Date: Wed, 22 Aug 2012 10:14:07 +0800 > To: <user@hive.apache.org> > Subject: hive comment in Chinese > > > > Hi, all > > > I tried to use Chinese in hive column comment and table comment,and > the metadata in Mysql is regular,the charset of 'COMMENT' column in > 'columns_v2' table and 'PARAM_VALUE' column in 'table_params' table both are > 'utf8'. > When I exec 'select * from columns_v2' with mysql client,the Chinese > comments display normally. But when I execute 'describe table' with hive > cli,the Chinese words are garbled.Why is it so? Please help me out. > > > > > Thanks & Regards > Zhan > >
