Skip to content

Data Types

The data for a process is described up front. The data description can be easily evolved over the process life. The VOR Stream system supports the following data types:

  • num (double precision floating point)
  • date
  • datetime
  • char (variable length strings - no practical limit to length)
  • array (double precision floating point)
  • bool – e.g., true, false, yes, no, 1, 0
  • int

Tip

Need to convert data types? See the data type conversions page for documentation on using the types package to convert between data types.

The following table details variable types:

Type Description Example CSV form
num This is a double precision floating point (64 bit) value. 1.5e-15, 22, NaN
char This is a variable length string of reasonable length. "I am a character string"
bool A Boolean variable is either true or false. T, F, true, false, 0, 1
date A date variable represented as days since the Epoch. In input files, dates should be in ISO 8601 or RFC 3339 date format (YYYY-MM-DD) although many other formats are parsed. The sequence is always day, month, then year or year, month, day. 1985/04/01, 2020-01-01, 01mar1970
datetime This is a datetime variable. Times are assumed to be UTC unless otherwise specified. 1994-11-05T13:15:30Z
int This is a 64 bit signed integer. -32,764
array There are two kinds; static and dynamic. A static array is a double precision floating point array of fixed length – arraylen. In input and output tables they are represented by sequentially named variable with numeric suffixes starting from 0 to arraylen-1. A dynamic array is a double precision floating point array of dynamic length. arraylen is set to -1. In input and output tables they are represented by a single column with comma separated values enclosed in quotes. myarry0, myarry1, myarry3 or 3.1415, 63, -200.1 , "10.5,9.5,4.5,3.5,0,1,2,3,4,5,6,7,8,9"

Missing values

VOR Stream represents missing data with a value specific to each type. For these common types, blank input cells and values that fail to parse become:

  • num → NaN
  • char → empty string
  • date → 1970-01-01

A blank num cell is accepted as NaN; a cell containing 0 stays zero. Test for a missing value with isnull(x) and supply a default with coalesce(...). VOR Stream has no NULL value. SQL aggregates treat missing the way standard SQL treats NULL: SUM, AVG, MIN, and MAX skip NaN inputs, COUNT counts non-missing values, and an all-missing group aggregates to NaN for SUM/MIN/MAX (AVG returns 0.0, COUNT returns 0). See the SQL node reference for details.