EVD and Data Types
‘EVD’ stands for ‘EVL Data Definition’ and it is the way how to specify structure of data sets in EVL. It can be used either inline, as a component option, or in an *.evd file.
EVL uses mostly standard C++ data types, so most of them are well known.
EVD Structure
Firstly an example of evd file, which defines data types for some csv file:
ID int sep=";"
Name string sep=";" null=""
"Birth Date" date("%-d.%-m.%Y") sep=";" null="1.1.1970"
Amount decimal(12,3) sep=";" thousands_sep=","
"Created At" datetime sep="\n" null="0000-00-00 00:00:00"
In general each nonempty line of EVD file looks like this:
<indent> Field_Name <blank> Data_Type <blank> EVD_Options |
where
- ‘<indent>’
might be empty, 2 spaces, 4 spaces, 6 spaces, etc., to define a substructure of compound data types, see Compound Types for details,
- ‘Field_Name’
is a sequence of any printable ASCII characters below 128. When a space is used, then whole field name must be quoted by double quotes. Special characters (also only ASCII ones under 128) must be escaped, e.g. ‘\n’, ‘\r’, ‘\t’, ‘\v’, ‘\b’, ‘\f’, ‘\a’, ‘\"’, ‘\\’, or in hexa ‘\x??’. Characters other than letters, numbers and underscore are replaced by underscore in mappings. All these field names are valid:
// Name in mapping: recommended_field_name // recommended_field_name "Field with a Space" // Field_with_a_Space 'field-with-a-hyphen' // _field_with_a_hyphen_ "$field_with_dollar" // _field_with_dollar 'single_quoted'field' // _single_quoted_field_ "with\nnewline" // with_newline
- ‘Data_Type’
is one of:
-
string, see String -
char,uchar,short,ushort,int,uint,long,ulong,int128,uint128, see Integral Types -
decimal, see Decimal -
float,double, see Float and Double -
date,time,time_ns,interval,interval_ns,datetime,timestamp, see Date and Time
-
- ‘EVD_Options’
is
<blank>separated list of options, see EVD Options- ‘<blank>’
is one or more spaces and/or tabs.
Comments
Standard C-style comments can be used in evd file, for example:
street_id int street_name string street_code string null="" // but NOT NULL in DB /* COMBAK: street_code will be replaced by street_num later this year street_num long */
Inline EVD
For the most of the EVL Components an inline EVD can be specified as an option. In such case comments are not allowed and the format is simply the same as for EVD in a file, just instead of newlines, commas are used to separate each field definition.
The same structure, as in above EVD Example, but as a component option (a comma separated list of fields with data types and options):
--data-definition='id int sep=";",
name string sep=";" null="",
birth_date date sep=";" null="1970-01-01",
amount decimal(12,3) sep=";" thousands_sep=",",
created_at datetime sep="\n" null="0000-00-00 00:00:00"'