Aaron McCurry created BLUR-237:
----------------------------------
Summary: Investigate the use of Thrift Unions to remove some
complexity in the Blur API
Key: BLUR-237
URL: https://issues.apache.org/jira/browse/BLUR-237
Project: Apache Blur
Issue Type: Improvement
Components: Blur
Affects Versions: 0.3.0
Reporter: Aaron McCurry
Fix For: 0.3.0
I believe that we can use this feature to provide a simpler API for Blur data
structures.
One place where this would be of benefit would be in changing the Column object
from having a string name and a string value. To having a string name and a
new Union type as the value. This ColumnValue union could contain one of each
of the primary thrift types (string, i32, i64, binary, double, etc). This
would allow for the original data type to be delivered in the correct form
(instead of just a string). It could also make the assignment of new types in
none strict tables to be easier. For example if the Union had a integer
provided and the column had not been seen before, then the type could be
assigned as an integer instead of using whatever the default type on the table
(TableDescriptor).
Another place is in the results of a query and fetch call. Currently you can
get either a Record or a Row back so I think a union type of the result would
work better here.
Union references:
https://issues.apache.org/jira/browse/THRIFT-409
http://web.archive.org/web/20090829121249/http://blog.rapleaf.com/dev/?p=48
http://web.archive.org/web/20100725051856/http://blog.rapleaf.com/dev/2009/08/18/thrift-unions-part-ii-or-how-i-reduced-memory-usage-by-95/
http://nathanmarz.com/blog/thrift-graphs-strong-flexible-schemas-on-hadoop.html
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
