There is no way to go around that right now. Note that the partition column is actually not in the data - it's appended to each row at runtime.
Hive uses the directory structure to figure out the partition values. On Mon, Apr 5, 2010 at 9:24 PM, Fan Yang <[email protected]> wrote: > I’ve found out the root cause is not the data and table format. It’s because > of the name of file on hdfs. I used “.data” as the file name and load data > command ignore the file when its name started with ‘.’ > > > > Anyway, thanks very much. > > > > BTW, when create a table with partitions. It seems the partition column must > be the last column of the table. How can I use some other column which is > not the last one to mark partition? > > > > Fan > > > > From: zhu weimin [mailto:[email protected]] > Sent: Tuesday, April 06, 2010 10:53 AM > To: [email protected] > Subject: RE: Problems of loading data from hdfs > > > > I think your data probably terminated by comma character. > > If is that, recreate table like the following > > CREATE TABLE table_name{ > > … > > } > > ROW FORMAT DELIMITED > > FIELDS TERMINATED BY ','; > > > > If is not, review your table describe and data samples > > > > Zhu weimin > > > > From: Fan Yang [mailto:[email protected]] > Sent: Tuesday, April 06, 2010 11:19 AM > To: [email protected] > Subject: Problems of loading data from hdfs > > > > Hi > > > > When I try to load data into table from hdfs, it seems successful and gives > no warnings. But using select clause can’t get any result from the table. > > However, after I load data from local file which is download from hdfs by > using ‘hadoop dfs get’ command, I can select out records. > > Note, these two files are exactly the same and no errors were given during > the procedure. > > > > Thanks in advance. > > > > Fan -- Yours, Zheng http://www.linkedin.com/in/zshao
