Hi, Instead of passing it as command line argument, I am storing all of this configuration in config/config.xml.
My earlier version was command line, but than as arguments increased I shifted to config.xml. Plus I thought providing all credentials at command line is also not a good idea. Sample Config file is https://github.com/nehalmehta/CSV2Cassandra/blob/master/config/config.xml. I am going to add following features: Cassandra Credentials, Selected Columns and selected primary key. I believe it is good idea to have function calls , which can manipulate selected csv columns before inserting records. Thanks, Nehal Mehta. On Tue, Sep 27, 2011 at 8:03 PM, Radim Kolar <h...@sendmail.cz> wrote: > > I have cleaned up my code that imports CSV into Cassandra and I have put > it open on > https://github.com/nehalmehta/**CSV2Cassandra<https://github.com/nehalmehta/CSV2Cassandra>. > Have a look if it is useful to you. > Hello, > I will remake this tool into something which is like Oracle SQL*Loader. > Basically, you will pass controlfile as command line argument. I need > conversion from DATE to milliseconds based date, header less CSV and better > CSV escaping. > > example of control file > > options (rows=1000) > LOAD DATA > INFILE 'c:\tmp\searches.csv' > BADFILE 'c:\tmp\searches.bad' > REPLACE > INTO TABLE SEARCHES2 > FIELDS TERMINATED BY "," > OPTIONALLY ENCLOSED BY '"' > ( query, > day date 'YYYY-MM-DD', > results, > ip > ) > > or maybe i will start project from 0 >
