Improved header handling

• Significantly improved support of table headers by using the enhanced query parsing algorithm.

Improvements and bug fixes

0.19.0

bug fix

compile & exec

• Python version uses compile + exec to execute generated python code instead of writing a temporary module

Named column variables for CSV

• Added support for 3 new variable notations (see #6):

1. attribute syntax, e.g. a.good_alphanumeric_column_name
2. dictionary syntax, e.g. a["arbitrary column name!"]
3. array syntax, e.g. a[3]

• Switched default encoding for CSV files to utf-8
• Implemented Join on NR, see (#10)
• Bug fixes

Better aggregate functions

• Added lowercase aliases for aggregate functions
• Improved error handling for aggregate functions
• Renamed "FOLD" -> "ARRAY_AGG"
• Renamed "UNFOLD" -> "UNNEST"
• Fixed bug with handling more than 10 strings in a query
• Unicode separators support by @veysiertekin

RFC-4180 compatibility for CSV parser

• CSV parser now supports RFC-4180 compatible newlines in fields via "quoted_rfc" dialect.
• Improved command line interface of rbql-js and rbql-py tools.