我正在使用一个相当大的数据集（大约500mio三元组），它存储在graphdb free中，并在本地开发人员机器上运行。
我想用rdf4j对数据集进行一些操作，并且必须或多或少地选择整个数据集。要进行测试，我只需选择所需的元组。代码在第一个一百万元组中运行良好，之后由于graphdb继续分配更多的ram，代码运行速度变得非常慢。
有没有可能对非常大的数据集执行select查询并批量获取它们？
基本上我只想通过一些选定的三元组进行“迭代”，所以不需要使用graphdb中的那么多ram。我可以看到，在查询完成之前，我已经准备好在rdf4j中获取数据，因为它只在大约140万个读取元组时崩溃（heapspaceerror）。不幸的是graphdb并没有释放allready读取元组的内存。我错过什么了吗？
非常感谢你的帮助。
ps.i allready将graphdb的可用堆空间设置为20gb。
rdf4j（java）代码如下所示：

package ch.test;

import org.eclipse.rdf4j.query.*;
import org.eclipse.rdf4j.repository.RepositoryConnection;
import org.eclipse.rdf4j.repository.http.HTTPRepository;

import java.io.File;
import java.io.IOException;
import java.nio.charset.StandardCharsets;
import java.nio.file.Files;
import java.nio.file.Paths;

public class RDF2RDF {

    public static void main(String[] args) {
        System.out.println("Running RDF2RDF");

        HTTPRepository sourceRepo = new HTTPRepository("http://localhost:7200/repositories/datatraining");
        try {
            String path = new File("").getAbsolutePath();
            String sparqlCommand= Files.readString(Paths.get(path + "/src/main/resources/sparql/select.sparql"), StandardCharsets.ISO_8859_1);

            int chunkSize = 10000;
            int positionInChunk = 0;
            long loadedTuples = 0;

            RepositoryConnection sourceConnection = sourceRepo.getConnection();
            TupleQuery query = sourceConnection.prepareTupleQuery(sparqlCommand);

            try (TupleQueryResult result = query.evaluate()) {
                for (BindingSet solution:result) {
                    loadedTuples++;
                    positionInChunk++;

                    if (positionInChunk >= chunkSize) {
                        System.out.println("Got " + loadedTuples + " Tuples");
                        positionInChunk = 0;
                    }
                }
            }

        } catch (IOException err) {
            err.printStackTrace();
        }
    }
}

选择.sparql：

PREFIX XXX_meta_schema: <http://schema.XXX.ch/meta/>
PREFIX XXX_post_schema: <http://schema.XXX.ch/post/>
PREFIX XXX_post_tech_schema: <http://schema.XXX.ch/post/tech/>

PREFIX XXX_geo_schema: <http://schema.XXX.ch/geo/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX XXX_raw_schema: <http://schema.XXX.ch/raw/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>

SELECT * WHERE {

    BIND(<http://data.XXX.ch/raw/Table/XXX.csv> as ?table).

    ?row XXX_raw_schema:isDefinedBy ?table.

    ?cellStreetAdress XXX_raw_schema:isDefinedBy ?row;
        XXX_raw_schema:ofColumn <http://data.XXX.ch/raw/Column/Objektadresse>;
        rdf:value ?valueStreetAdress.

    ?cellOrt mobi_raw_schema:isDefinedBy ?row;
        XXX_raw_schema:ofColumn <http://XXX.mobi.ch/raw/Column/Ort>;
        rdf:value ?valueOrt.

    ?cellPlz mobi_raw_schema:isDefinedBy ?row;
        XXX_raw_schema:ofColumn <http://XXX.mobi.ch/raw/Column/PLZ>;
        rdf:value ?valuePLZ.

    BIND (URI(concat("http://data.XXX.ch/post/tech/Adress/", MD5(STR(?cellStreetAdress)))) as ?iri_tech_Adress).
}

我的解决方案是：使用subselect语句，它首先获取所有“行”。

PREFIX mobi_post_schema: <http://schema.mobi.ch/post/>
PREFIX mobi_post_tech_schema: <http://schema.mobi.ch/post/tech/>

PREFIX mobi_geo_schema: <http://schema.mobi.ch/geo/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX mobi_raw_schema: <http://schema.mobi.ch/raw/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>

SELECT * WHERE {

    {
        SELECT ?row WHERE
        {
            BIND(<http://data.mobi.ch/raw/Table/Gebaeudeobjekte_August2020_ARA_Post.csv> as ?table).

            ?row mobi_raw_schema:isDefinedBy ?table.
        }
    }

    ?cellStreetAdress mobi_raw_schema:isDefinedBy ?row;
        mobi_raw_schema:ofColumn <http://data.mobi.ch/raw/Column/Objektadresse>;
        rdf:value ?valueStreetAdress.

    ?cellOrt mobi_raw_schema:isDefinedBy ?row;
        mobi_raw_schema:ofColumn <http://data.mobi.ch/raw/Column/Ort>;
        rdf:value ?valueOrt.

    ?cellPlz mobi_raw_schema:isDefinedBy ?row;
        mobi_raw_schema:ofColumn <http://data.mobi.ch/raw/Column/PLZ>;
        rdf:value ?valuePLZ.

    BIND (URI(concat("http://data.mobi.ch/post/tech/Adress/", MD5(STR(?cellStreetAdress)))) as ?iri_tech_Adress).
}