[DRILL-261] Implement support for DECIMAL type

• Type: Task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.4.0
• Assignee: Unassigned
• Reporter: Ben Becker
• Created: 10/Oct/13 18:46
• Updated: 31/Jul/14 06:31
• Resolved: 29/May/14 15:44

Description

Many functions and operators in Drill need to handle decimal types. This may mean two formats; an expanded format for use during processing, and a compact format for efficient representation on disk.

Postgres' compact (on-disk) format for the decimal types NumericShort and NumericLong generally fit into a 64 bits:

1 bit for sign
6 bits for scale  in NumericShort; 16 bits in NumericLong
7 bits for weight in NumericShort; 16 bits in NumericLong
16 bits for the digits

The expanded (in-memory) format in Postgres is defined in numeric.c as follows:

typedef struct NumericVar
{
	int ndigits;	/* # of digits in digits[] - can be 0! */
	int weight;	/* weight of first digit */
	int sign;	/* NUMERIC_POS, NUMERIC_NEG, or NUMERIC_NAN */
	int dscale;	/* display scale */
	NumericDigit *buf;      /* start of palloc'd space for digits[] */
	NumericDigit *digits;	/* base-NBASE digits */
} NumericVar;

Additional notes and links to related published papers are in the aforementioned source file.