Skip to end of metadata
Comment: Migrated to Confluence 4.0
Go to start of metadata

This page is used to generate the BNF grammar for the new proposed TXT format (an extension of CQL).

To generate a stable BNF we are going to do the following:

  1. some examples (this is going to be the a user document), to reduce the work
    we could add only, examples of cql changes and new (no presents in cql).
  2. for each example we will write a grammar rule in BNF (important: maintain
    reference example -> BNF rule).
  3. If we have complete grammar, we could begin to: eliminate left recursions,
    ambiguous rules analysis, etc.
  4. programming: unit test (from cql and 1 ); write the javacc input (from 3);
    refactor cql to reuse the builders.

This page is being used to record the notes for step 1 through 3...

Related

Contents

Examples of Modified Rules

Basically we are going to replace the use of attribute_name with expression except where it does not make sense:

  • exists check is only valid for attribute_name

Comparisons

CQL Example:

CQL Rule:

<comparison predicate> ::=
    <attribute_name> <comp op> <literal>
    | <attribute_name> <comp op> <expression>

TXT Examples:

TXT Rule:

<comparison predicate> ::= <expression> <comp op> <expression>

Between

CQL Examples:

CQL Rule:

<between predicate> ::= <attribute-name> [ NOT ] BETWEEN <literal> AND <literal>

TXT Examples:

TXT Rule:

<between predicate> ::= <expression> [ "NOT" ] "BETWEEN" <expression> "AND" <expression>

LIKE Text Pattern

CQL Example:

foo LIKE 'John%'

CQL Rule:

<text predicate> ::= <attribute name> [ "NOT" ] "LIKE" <character pattern>

TXT Example:

strconcat("John", "Oliver") LIKE 'John%'

TXT Rule:

<like text predicate> ::= <expression> [ "NOT" ] "LIKE" <character pattern>

Id Predicate

CQL Example: n/a

TXT Examples:

 ID IN 'river.43'
 ID IN 'road.1', 'road.2', 'road.47'

TXT Rule:

<id predicate> ::= "ID" ["NOT"] "IN" "(" <id> {"," <id> } ")"
<id> ::=

IN predicate

CQL Example: n/a

TXT Examples:

 huc_8 IN (4100001,4100011,4100012)
 huc_8 NOT IN (4100001,4100011,4100012)

TXT Rule:

<in predicate>    	::=   <attribute-name> [  "NOT"  ]  "IN"  <in predicate value>
<in predicate value>    ::=   "(" <in value list> ")"

Null

CQL Examples:

ATTR1 IS NULL
ATTR2 IS NOT NULL

CQL Rule:

<null predicate> ::= <attribute name> "IS" [ "NOT" ] "NULL"

TXT Examples:

ATTR1 IS NULL
ATTR2 IS NOT NULL
centroid( the_geom ) IS NOT NULL

TXT Rule:

<null predicate> ::= <expression> "IS" [ "NOT" ] "NULL"

Spatial Predicates Functions

The spatial predicates are expressed in CQL with function syntax as the following examples.

CQL Example:

CONTAINS(the_geom  POINT(12 30))
BBOX(ATTR1, 10,20,30,40)

CQL Rule:

<routine invocation> ::=
          <geoop name><georoutine argument list>
	| <relgeoop name><relgeoop argument list>
	| <routine name><argument list>
	| "BBOX" <bbox argument list>

This extension allows the composition of functions

TXT Example:

CONTAINS(the_geom, POINT(12 30))

BBOX(ATTR1, 10,20,30,40)

CROSSES(centroid( the_geom ), BUFFER(LINESTRING (15 15, 20 20),10))

BBOX( buffer( the_geom , 10), 10,20,30,40 )
BBOX( buffer( the_geom , 10), buffer( POINT( 15,15), 10) )

TXT Rule:

The name of spatial relations were adjusted to Implementation Specification for Geographic Information - Simple Feature Access ( OGC 05-134 v1.1.0 ).

<geometry predicate> ::= <geoop name><georoutine argument list>
	| <expression><relgeoop name><relgeoop argument list>
	| BBOX <bbox argument list>

<geoop name> ::=   EQUALS | DISJOINT | INTERSECTS | TOUCHES | CROSSES | WITHIN | CONTAINS |OVERLAPS | RELATE
<georoutine argument list> ::=  <expression> | <geometry literal>

<relgeoop name> ::= "DWITHIN" | "BEYOND"
<relgeoop argument list> ::= <expression> <tolerance>
	| <geometry literal> <tolerance>

bbox argument list> ::= "(" <attribute>,<min X> ,<min Y> ,<max X>,<max Y> ["," <crs>] ")"
	| "(" <expression>,<min X> ,<min Y> ,<max X>,<max Y> ["," <crs>] ")"
        | "(" <expression>, <expression> ")"

Spatial Relation Predicate

Thinking to improve the legibility and achieve a language near to "natural language" in this section is proposed an alternative syntax to the "Spatial Predicates Functions".

CQL Example: see "Spatial Geometry Predicates" section

TXT Example:

geom1 INTERSECTS geom2

geom1 TOUCHES geom2

geom1 EQUALS geom2

geom1 DISJOINTS geom2 

geom1 CROSSES geom2 

geom1 WITHIN geom2 

geom1 CONTAINS geom2

geom1 OVERLAPS geom2 

geom1 RELATES geom2 LIKE 'T**F*****'

geom1 DWITHIN geom2  DISTANCE 10.0 km 

geom1 BEYOND geom2  DISTANCE 10.0 km

TXT Rule:

<spatial-relation predicate> ::= 
      <geom-expression>  <spatial-relation> <geom-expression>
     |<geom-expression> <distance-relation> <geom-expression>  DISTANCE <number> <units>
     |<geom-expression> "RELATES" <geom-expression>  [ "NOT" ] "LIKE" <relate pattern>


<spatial-relation> ::= EQUALS | DISJOINTS | INTERSECTS | TOUCHES 
                     | CROSSES | WITHIN | CONTAINS |OVERLAPS | RELATES

<distance-relation> ::=  DWITHIN | BEYOND

<relate pattern> ::=  <quote>
                         <relate flag><relate flag><relate flag><relate flag>
                         <relate flag><relate flag><relate flag><relate flag>
                         <relate flag>
                        <quote>

<relate flag> ::= "T" | "F" | "*" | "0" | "1" | "2"

TXT Grammar

We provide a new TXT grammar which includes the described syntax rules in the following link:TXT BNF Design

Labels
  • None