Literals (OLAP)

[Table of Contents](https://github.com/dell-oss/Doradus/wiki/OLAP Databases: Table-of-Contents) | Previous | Next
Doradus Query Language: Literals

---

Most clauses contain a comparison that include a literal. When the comparison uses a predefined field whose type is known, the literal _should_ use a format compatible with that type. However, Doradus does not strictly enforce this. For example, if `Size` is declared as an integer, the following clause is allowed:

Size=Foo

Of course, because Size is declared as an integer, it cannot match the value Foo, so no objects will be selected by the clause.

The types of literals supported by Doradus are described below.

Booleans

A Boolean literal is the keyword true or false (case-insensitive).

Numbers

Doradus supports both integer and floating-point numeric literals. An integer literal is all digits, preceded with a minus sign (-) for negative values. Integer literals are considered long (64-bit). Floating-point constants use the same format supported by Java: an optional preceding sign, at least one digit, an optional fractional part, and an optional exponent, which may be signed. Example floating-point literals are shown below:

0
-123
3.14
10E12
2.718e-6  

Timestamps

Timestamp literals have the general form:

"YYYY-MM-DD hh:mm:ss.fff"

Notes:

• In DQL expressions, timestamp values should be enclosed in single or double quotes.
• All time and date elements except for the year can be omitted from right-to-left if the preceding separator character ('.', ':', or '-') is also eliminated.
• When a timestamp value consists of a year only, the enclosing quotes can be omitted.
• Omitted time components default to 0; omitted date components default to 1.

Terms

Text fields can be evaluated as a set of tokens called terms. Depending on the underlying storage service, the terms may be generated and indexed when a text field is stored, or they may be computed dynamically. Generally, terms are alphanumeric character sequences separated by whitespace and/or punctuation.

In search clauses, multiple terms can be used, and each term can contain wildcards. Example search terms are shown below:

John
Fo*
a?c
Event_413  

Note that terms are not quoted. A term is a sequence of characters consisting of letters, digits, or any of these characters:

? (question mark)
* (asterisk)
_ (underscore)
@ (at symbol)
# (hash or pound)
- (dash or minus)  

In a term, the characters _, @, #, and - are treated as term separators, creating multiple terms that are part of the same literal. For example, the following literals all define two terms, john and smith – the separator character is not part of either term:

john_smith
john@smith
john#smith
john-smith  

The characters ? and * are single- and multi-character wildcards: ? matches any single character and * matches any sequence of characters.

See the discussion on Text Contains Clauses for more details about using terms in searches.

Strings

Some clauses compare to a full text string instead of a term. When the text string conforms to the syntax of a term, it can be provided unquoted. Example:

Name=Smith

The value Smith is a string because "=" follows the field name; if the field name was followed by ":", Smith would be treated as a term. See Equality Clauses described later.

When the string consists of multiple terms or contains characters not allowed in terms, the string must be enclosed in single or double quotes. Example:

Name="John Smith"

Note that to Doradus, object IDs are also strings. Therefore, when used in DQL queries, object ID literals may be provided as a single unquoted term or as a quoted string. Examples:

_ID=ISBN1234
Sender='cihSptpZrCM6oXaVQH6dwA=='  

In a string, the characters ? and * are treated as wildcards as they are in terms. If a string needs a ? or * that must be used literally, they must be escaped. Escaping can also be used for non-printable or other characters. Doradus uses the backslash for escaping, and there are two escape formats:

• \x: Where x can be one of these characters: t (tab character), b (backspace), n (newline or LF), r (carriage return), f (form feed), ' (single quote), " (double quote), or \ (backslash).
• \uNNNN: This escape sequence can be used for any Unicode character. NNNN must consist of four hex characters 0-F.

For example, the string "Hello!World" can also be specified as "Hello\u0021World".