我将第二分割这一个，进一次创建节点和创建关系（每个）：
USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM "file:///Users/btibert/Dropbox/Projects/bentley-search-neo4j/data/templates.csv" AS row
WITH row
MATCH (r:Vendor {name:row.vendor})
WITH row, r
MERGE (p:Template {name:row.template_clean})
MERGE (v:Version {version:row.template_ver})
MERGE (p)-[:FROM_VERSION]->(v)
MERGE (p)-[:CREATED_BY]->(r);
正如你可以清楚地看到在计划中渴望操作。
我的意思是，如果你只有几千行这并不重要。但是，如果它走向数十万甚至数百万，那么拉动所有数据需要更多的内存。如果你有每个学生可以使用WITH DISTINCT toInt(row.pidm) as pidm, ....数量减少的合并它运行多行同时 ：
+----------------+------------------------------------+------------------------------------------------------------------------------------------------+
| Operator | Identifiers | Other |
+----------------+------------------------------------+------------------------------------------------------------------------------------------------+
| EmptyResult | | |
| UpdateGraph(0) | anon[270], anon[301], p, r, row, v | MergePattern |
| UpdateGraph(1) | anon[270], p, r, row, v | MergePattern |
| UpdateGraph(2) | p, r, row, v | MergeNode; row.template_clean; :Template(name); MergeNode; row.template_ver; :Version(version) |
| Eager | r, row | |
| SchemaIndex | r, row | row.vendor; :Vendor(name) |
| LoadCSV | row | |
+----------------+------------------------------------+------------------------------------------------------------------------------------------------+
我可能会改变这种为ON CREATE SET变种对于非关键属性。
LOAD CSV WITH HEADERS FROM "recs.csv" AS row
WITH row
MERGE (s:Student {pidm:toInt(row.pidm)})
ON CREATE SET s.hash_pidm=toInt(row.hash_pidm), ....;
这一次我会分成两个语句，每一个关系，否则你可能会得到太多的比赛： （你不需要WITH S IN之间）
LOAD CSV WITH HEADERS FROM "...recs.csv" AS row
WITH row
MATCH (s:Student {pidm: toInt(row.pidm)})
MATCH (v:Vendor {name: row.vendor})
MATCH (a:Ability {name: row.ability})
WITH row, s, v, a
MERGE (s)-[:PURCHASED_FROM]->(v)
MERGE (s)-[:HAS_ABILITY]->(a);
将成为：
LOAD CSV WITH HEADERS FROM "...recs.csv" AS row
MATCH (s:Student {pidm: toInt(row.pidm)})
MATCH (v:Vendor {name: row.vendor})
MERGE (s)-[:PURCHASED_FROM]->(v);
LOAD CSV WITH HEADERS FROM "...recs.csv" AS row
MATCH (s:Student {pidm: toInt(row.pidm)})
MATCH (a:Ability {name: row.ability})
MERGE (s)-[:HAS_ABILITY]->(a);
这里我还要对自己创建的联系人信息。 （再次上创建SET） ，做在一个单独的语句中的师生关系：
LOAD CSV WITH HEADERS FROM "....cont.csv" AS row
MERGE (c:Contact {cid:row.cid}) ON CREATE SET ....;
LOAD CSV WITH HEADERS FROM "...cont.csv" AS row
MATCH (s:Student {pidm:toInt(row.pidm)})
MATCH (c:Contact {cid:row.cid})
MERGE (s)-[:HAS_CONTACT]->(c);
我还要拆这一个到两个语句：
LOAD CSV WITH HEADERS FROM "...cont.csv" AS row
WITH row WHERE toInt(row.seqnum) = 1
MATCH (s:Student {pidm:toInt(row.pidm)})
MATCH (f:Contact {cid:row.first_cont})
MERGE (s)-[:FIRST]->(f);
LOAD CSV WITH HEADERS FROM "...cont.csv" AS row
WITH row WHERE toInt(row.seqnum) = 1
MATCH (s:Student {pidm:toInt(row.pidm)})
MATCH (l:Contact {cid:row.last_cont})
MERGE (s)-[:LAST]->(l);
斯普利特这一个进入E-邮件创建和后来通过MSG-ID将其连接到学生：
LOAD CSV WITH HEADERS FROM "...brm.csv" AS row
MERGE (e:Email {msgid:row.msgid}) ON CREATE SET ... ;
LOAD CSV WITH HEADERS FROM "file:///Users/btibert/Dropbox/Projects/bentley-search-neo4j/data/brm.csv" AS row
MATCH (s:Student {pidm:toInt(row.pidm)})
MATCH (e:Email {msgid:row.msgid})
MERGE (s)-[:WAS_SENT]->(e);
HTH迈克尔
在您最近的加载CSV语句中，您使用'Email：cid'作为属性，而您的约束位于'电子邮件：msgid' –
好，赶快看看结果如何。 – Btibert3
这次，当我尝试从cypher.cql文件的第111行开始执行命令时，浏览器中出现“数据库断开连接”警告。从命令行运行该文件是否会提供任何性能改进？ – Btibert3