Definitions for Values Insertion #177

afs · 2025-01-08T10:53:48Z

A replacement for substitute.

The context is taken from SEP-0007.

Currently, all new material is in a single section.

I am submitting this PR to put the content under the WG agreement for IP and Licensing.

Preview | Diff

spec/index.html

TallTed · 2025-01-08T18:14:26Z

spec/index.html

+                This applies to `BIND`, variables introduced by `AS` in a `SELECT` clause, variables in a `VALUES`
+                clause, and variables intrucded by `AS` in `GROUP BY`.
+              
+              Extend the "in-scope" rules to include the variables in-scope from the current row:


Suggested change

<p>Extend the "in-scope" rules to include the variables in-scope from the current row:p>

<p>Extend the "in-scope" rules to include the variables that are in-scope for the current row:p>

spec/index.html

TallTed · 2025-01-08T18:23:28Z

spec/index.html

+
+              

            
+                A variable inside a project expression that is not in the variables projected
+                is not affected by the values insertion operation because it is renamed apart.


Suggested change

is not affected by the values insertion operation because it is renamed apart.

is not affected by the value insertion operation because it is renamed apart.

TallTed · 2025-01-08T18:24:42Z

spec/index.html

+          
+          

            
+            Values Insertion and `EXISTS`


Values Insertion reads very oddly to me. Can this be changed (throughout) to Value Insertion?

Suggested change

<section>

<h3>Values Insertion and `EXISTS`h3>

<section>

<h3>Value Insertion and `EXISTS`h3>

It is inserting multiple values.

Do you use an "eggs beater" or an "egg beater"?

In US English, if not global English, the common phrasing for such things uses the singular — "rule book" (not "rules book"), "tool box" (not "tools box"), "book shelf" (not "books shelf").

Also, is this only and always "inserting multiple values", or can it be "inserting a single value"?

TallTed · 2025-01-08T18:26:44Z

spec/index.html

+                

            
+                  Definition: Values Insertion
+                  

            
+                    Define the Values Insertion function `ValuesInsert(X, μ)`


If, as I suggested above, Values Insertion is being changed to Value Insertion, then ValuesInsert should likewise be changed to ValueInsert

Suggested change

Define the Values Insertion function `ValuesInsert(X, μ)`

Define the Value Insertion function `ValueInsert(X, μ)`

While this is in algebra space, the syntax analogy is VALUES.

The μ can have several bindings, correlating several variables.

Why not simply InsertValues ?

TallTed · 2025-01-08T18:28:42Z

spec/index.html

+                exists(X) = true 

            
+              if eval( D(G), ValuesInsert(PrjMap(X), μ) )
+                 is a non-empty solution sequence.
+exists(X) = false otherwise


Suggested change

exists(X) = false otherwisepre>

Otherwise, exists(X) = false.pre>

The style here is that for function definition, with each case being aligned. It is not running text.

pfps · 2025-05-19T10:23:30Z

In my view "values insertion" is the correct phrase to use.

hartig · 2025-05-23T08:59:36Z

spec/index.html

+                  a partial mapping `F` from
+                  `V`,
+                  the set of all variables, to `V` where:
+                
+                F(v) = v1 if v is in PV, where v1 is a fresh variable

            
+F(v) = v if v is not in PV


Considering the second line of this definition, it seems to me that F is actually not partial.

If I am wrong with this assumption, can you explain which variables are mapped by F and which ones are not.

hartig · 2025-05-23T09:41:45Z

spec/index.html

+                  The Projection Expression Variable Remapping yields an algrebra expression that
+                  evaluates to the same results as the Project argument. No variable of `ProjectMap(Project(A, PV))`
+                  that is not in `PV` is mentioned anywhere else in the algebra expression for the query.


I don't think that's true based on the current version of this definition. By the current definition, the projection variables are renamed from A to A1, but not in PV. As a consequence, the evaluation of A1 will not even produce bindings for the projection variables.

Is it possible that you have a mistake in the earlier part of this definition? In particular, I have the feeling that the first line of the definition of F should have been as follows.

F(v) = v1 if v is mentioned in A but is not in PV , where v1 is a fresh variable

hartig · 2025-05-23T11:59:30Z

spec/index.html

+                  PrjMap(X) = replace all project operations Project(P, PV) 

            
+                with ProjectMap(P, PV) for each projection in X.


Is the ProjectMap(P, PV) here actually meant to be ProjectMap( Project(P, PV) ) as introduced in the previous definition? If not, how is ProjectMap(P, PV) defined?

hartig · 2025-05-23T12:20:39Z

spec/index.html

+              if eval( D(G), ValuesInScope(PrjMap(X)), μ)
+                 is a non-empty solution sequence.


What is ValuesInScope( .. )? Where is this defined?

No matter how it is defined, there is another problem here. The formula uses the eval function with three arguments: i) D(G), ii) the output of ValuesInScope( .. ), whatever that might be, and iii) some solution mapping μ. Yet, eval is not defined for such arguments; it is defined for two arguments: i) D(G) and ii) an algebraic expression.

hartig · 2025-05-23T13:02:53Z

spec/index.html

+Replace each occurence of `Y` in X where `Y` is one of
+          Basic Graph Pattern,
+          Property Path Expression,
+          `Graph(Var, pattern)`,
+          Inline Data
+          with `join(Y, BindingInScope())`.


What is X?

Instead of Property Path Expression, do you actually mean Property Path Pattern?

pfps · 2025-05-23T14:03:04Z

I note that this PR is for deep values insertion where values are insterted into subqueries.

hartig · 2025-05-23T14:31:08Z

spec/index.html

+                    Define the function `BindingInScope()`. 
+                    Evaluation of `BindingInScope()` results in a table of one row,
+                    being the current binding of the enclosing filter.


EXIST can be used not only within FILTER.

Where else?

Bind, select expression, order by…

Added: https://github.com/afs/SPARQL-exists/tree/main/tests/exists-6-places

afs · 2025-05-30T08:03:28Z

I note that this PR is for deep values insertion where values are insterted into subqueries.

Only variables are "in-scope" are affected.
The renaming apart means that variables hidden by scope in subqueries are not correlated. This is a point where correlation and the requirements of parameterisation might differ.

This query has a sub-query that exposes ?o

SELECT * {
    ?s :p ?o .
    FILTER EXISTS{ SELECT ?o { :s :q ?o }}

whereas the inner ?o in

SELECT * {
    ?s :p ?o .
    FILTER EXISTS{ SELECT ?s { ?s :q :?o }}

is not correlated.

As algebra expressions:

SELECT * {
    ?s :p ?o .
    FILTER EXISTS { SELECT * { ?s :q ?z} }

SELECT * {
    ?s :p ?o .
    FILTER EXISTS { ?s :q :?z }

are the same.

afs changed the title ~~Defintions for Values Insertion~~ Definitions for Values Insertion Jan 8, 2025

afs force-pushed the values-insertion branch from 3f606dd to f837588 Compare January 8, 2025 10:58

TallTed suggested changes Jan 8, 2025

View reviewed changes

afs mentioned this pull request Jan 16, 2025

GH-2924: Lateral - fixed injection for tables and enhanced QueryExec API for easier testing. apache/jena#2925

Merged

3 tasks

afs self-assigned this Jan 23, 2025

afs force-pushed the values-insertion branch from f3902e7 to 7e6e82a Compare April 4, 2025 14:06

afs force-pushed the values-insertion branch 2 times, most recently from 0818131 to 95b4b0a Compare April 30, 2025 19:01

afs mentioned this pull request May 3, 2025

simple exists changes #211

Draft

Initial values-insertion draft

0f7b66d

afs force-pushed the values-insertion branch 3 times, most recently from c00cb1b to a0ccbd5 Compare May 18, 2025 19:41

Add 'row in scope' approach

f852c60

afs force-pushed the values-insertion branch from a0ccbd5 to f852c60 Compare May 18, 2025 19:42

afs mentioned this pull request May 19, 2025

Addressing SPARQL EXISTS errata #156

Open

hartig reviewed May 23, 2025

View reviewed changes

	<p>Extend the "in-scope" rules to include the variables in-scope from the current row:p>
	<p>Extend the "in-scope" rules to include the variables that are in-scope for the current row:p>



		A variable inside a project expression that is not in the variables projected
		is not affected by the values insertion operation because it is renamed apart.

	is not affected by the values insertion operation because it is renamed apart.
	is not affected by the value insertion operation because it is renamed apart.

-          <section>
-            <h3>Values Insertion and `EXISTS`h3>
+          <section>
+            <h3>Value Insertion and `EXISTS`h3>


		Definition: Values Insertion

		Define the Values Insertion function `ValuesInsert(X, μ)`

	Define the Values Insertion function `ValuesInsert(X, μ)`
	Define the Value Insertion function `ValueInsert(X, μ)`

	exists(X) = false otherwisepre>
	Otherwise, exists(X) = false.pre>

		PrjMap(X) = replace all project operations Project(P, PV)
		with ProjectMap(P, PV) for each projection in X.

		if eval( D(G), ValuesInScope(PrjMap(X)), μ)
		is a non-empty solution sequence.

Definitions for Values Insertion #177

Are you sure you want to change the base?

Definitions for Values Insertion #177

Uh oh!

Conversation

afs commented Jan 8, 2025 • edited by pr-preview bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Values Insertion and `EXISTS`

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pfps commented May 19, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pfps commented May 23, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

afs commented May 30, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

afs commented Jan 8, 2025 •

edited by pr-preview bot

Loading

afs commented May 30, 2025 •

edited

Loading