Flamy uses the following terms and conventions:
A schema represents a set of tables.
It is uniquely identified by its name (eg:
A table represents a Hive table.
Although Hive does not, Flamy requires users to always refer to a table with its fully qualified name (eg:
An item may represent either a Schema or a Table. When a command requires ITEMS as arguments, the user can specify any space-separated list of table and/or schema names. Giving a schema name is equivalent to giving each table names inside this schema.
A partition represents one partition of a Hive Table. If you are not familiar with partitioning in Hive, checkout this tutorial and try to use them, partitions are great! Just don’t try to have too many for one table… As a rule of thumb, tables with more than a few thousands partitions may start causing issues.
A partition is identified by a string of the form:
However, the ordering of the columns in the string do not matter for Flamy (even if it does for Hive and HDFS).
stats.daily_visitors/campaign=shoes/day=2014-10-12 works too)
/ being used as delimiters in the partitions names, they should not be used
as schema name, table name, partition key or value.