提问参考模版：
MATCH p = (self:Company)<-[sr:ACTUAL|CONTROL]-(target) WHERE self.keyNo = "6b58ef554c24961281b58f01cd676511" RETURN target
如何将上述语句改写为go 语句
注：target 有很多类型，如果用$$.Tag1.property1, $$.Tag2.property2, $$.Tag3.property3 这种方式 要写很多很多列
借助 properties() 函数
LOOKUP ON Company WHERE Company.keyNo == "6b58ef554c24961281b58f01cd676511" \
YIELD id(vertex) AS VertexID \
| GO FROM $-.VertexID OVER ACTUAL,CONTROL REVERSELY YIELD properties($^)
例子
(root@nebula) [basketballplayer]> LOOKUP ON player WHERE player.name == "Tim Duncan" YIELD id(vertex) AS VertexID| GO FROM $-.VertexID OVER follow,serve YIELD properties($$)
+----------------------------------+
| properties($$) |
+----------------------------------+
| {age: 36, name: "Tony Parker"} |
| {age: 41, name: "Manu Ginobili"} |
| {name: "Spurs"} |
+----------------------------------+
Got 3 rows (time spent 3834/29784 us)
这个地方LOOKUP 需要加索引了，目前线上服务没有加索引，可以理解vid 是对keyno 的hash，该怎么改写呢？
match/lookup 这里如果都是把 keyNo 作为属性都是需要索引的哈。
如果实际上 keyNo 是唯一的话，可以设计成 VID，就不需要索引，这样可以直接一个 GO from 就行了哈。
才看到是hash 的关系，假设您这个 hash 是nebula的这个的话，如果不是的话，就您在 应用层自己算 hash 替换这里。
GO FROM hash("6b58ef554c24961281b58f01cd676511") OVER ACTUAL,CONTROL REVERSELY YIELD properties($^)
https://docs.nebula-graph.com.cn/2.6.1/3.ngql-guide/6.functions-and-expressions/12.hash/
感谢，非常有用，如果这种呢？能不能改写，collect(p)，max(sr.percent) 我都没有想到较好的写法
MATCH
p = (self:Company)-[sr:Hold]->()<-[tr:Invest]-()
WHERE
self.keyNo = "6b58ef554c24961281b58f01cd676511"
AND tr.stockPercent >= 10
AND target.keyNo <> "6b58ef554c24961281b58f01cd676511"
RETURN
CASE target.status WHEN '存续' THEN 0 WHEN '在业' THEN 0 WHEN '' THEN 0 WHEN '迁入' THEN 1 WHEN '迁出' THEN 1 WHEN '停业' THEN 10 WHEN '清算' THEN 10 WHEN '吊销' THEN 11 WHEN '注销' THEN 11 ELSE 8 END AS status,
target,
collect(p) AS path,
'HE-BE10' AS nodeType,
max(sr.percent) AS p1,
max(tr.stockPercent) AS p2
ORDER BY
status, p1 desc, p2 desc
SKIP 0 LIMIT 10
这是我想到的改写方式，就很繁琐
go from -6732559126034379613 OVER Hold yield $^.Company.name as srcName, $^.Company.keyno as srcKeyno, $$.Company.name as mName, $$.Company.keyno as mKeyno, Hold._dst as id, Hold.stockpercent as sr, Hold.type as type| go from $-.id OVER Invest Reversely where Invest.stockpercent >= 10 and Invest._dst != -6732559126034379613 yield CASE $$.Company.shortstatus WHEN '存续' THEN 0 WHEN '在业' THEN 0 WHEN '' THEN 0 WHEN '迁入' THEN 1 WHEN '迁出' THEN 1 WHEN '停业' THEN 10 WHEN '清算' THEN 10 WHEN '吊销' THEN 11 WHEN '注销' THEN 11 ELSE 8 END AS status, $-.srcName as srcName, $-.srcKeyno as srcKeyno, $-.sr as relPercent, $-.type as relType, $-.mName as middleName, $-.mKeyno as middleKeyno, $$.Company.name as targetName, $$.Company.keyno as targetKeyno, Invest.role as rel2Role, Invest.type as rel2Type, Invest.stockpercent as rel2Percent, $-.sr as p1, Invest.stockpercent as p2 | ORDER BY status ASC, p1 desc, p2 desc | LIMIT 10
这种输出 path 的结果，用 GO 确实不太能表达，我假设你的query类似于这个
MATCH
p=(self:player)-[sr:serve]->()<-[tr:serve]-(target)
WHERE
id(self) == "player100"
AND tr.start_year >= 1995
RETURN
target.player.age AS status,
target,
collect(p) AS `path`,
'HE-BE10' AS nodeType,
max(sr.start_year) AS p1,
max(tr.start_year) AS p2
ORDER BY
status, p1 desc, p2 desc
SKIP 0 LIMIT 10
+--------+-----------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+------+------+
| status | target | path | nodeType | p1 | p2 |
+--------+-----------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+------+------+
| 25 | ("player106" :player{age: 25, name: "Kyle Anderson"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2018, start_year: 2014}]-("player106" :player{age: 25, name: "Kyle Anderson"})>] | "HE-BE10" | 1997 | 2014 |
| 27 | ("player110" :player{age: 27, name: "Cory Joseph"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2015, start_year: 2011}]-("player110" :player{age: 27, name: "Cory Joseph"})>] | "HE-BE10" | 1997 | 2011 |
| 29 | ("player113" :player{age: 29, name: "Dejounte Murray"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2019, start_year: 2016}]-("player113" :player{age: 29, name: "Dejounte Murray"})>] | "HE-BE10" | 1997 | 2016 |
| 29 | ("player112" :player{age: 29, name: "Jonathon Simmons"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2017, start_year: 2015}]-("player112" :player{age: 29, name: "Jonathon Simmons"})>] | "HE-BE10" | 1997 | 2015 |
| 31 | ("player105" :player{age: 31, name: "Danny Green"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2018, start_year: 2010}]-("player105" :player{age: 31, name: "Danny Green"})>] | "HE-BE10" | 1997 | 2010 |
| 32 | ("player104" :player{age: 32, name: "Marco Belinelli"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@20132015 {end_year: 2015, start_year: 2013}]-("player104" :player{age: 32, name: "Marco Belinelli"})>, <("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@20182019 {end_year: 2019, start_year: 2018}]-("player104" :player{age: 32, name: "Marco Belinelli"})>] | "HE-BE10" | 1997 | 2018 |
| 32 | ("player103" :player{age: 32, name: "Rudy Gay"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2019, start_year: 2017}]-("player103" :player{age: 32, name: "Rudy Gay"})>] | "HE-BE10" | 1997 | 2017 |
| 32 | ("player107" :player{age: 32, name: "Aron Baynes"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2015, start_year: 2013}]-("player107" :player{age: 32, name: "Aron Baynes"})>] | "HE-BE10" | 1997 | 2013 |
| 33 | ("player102" :player{age: 33, name: "LaMarcus Aldridge"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2019, start_year: 2015}]-("player102" :player{age: 33, name: "LaMarcus Aldridge"})>] | "HE-BE10" | 1997 | 2015 |
| 34 | ("player109" :player{age: 34, name: "Tiago Splitter"}) | [<("player100" :player{age: 42, name: "Tim Duncan"})-[:serve@0 {end_year: 2016, start_year: 1997}]->("team204" :team{name: "Spurs"})<-[:serve@0 {end_year: 2015, start_year: 2010}]-("player109" :player{age: 34, name: "Tiago Splitter"})>] | "HE-BE10" | 1997 | 2010 |
+--------+-----------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+------+------+
其实 collect 、max、min 的表达在 GO 里边是 GROUP BY … YIELD collect/max/min
我直接写出来是这样的
GO FROM "player100" OVER serve YIELD serve._dst AS id, serve.start_year AS sr_start_year, properties($$) AS middle_vertex, properties(edge) AS edge0, properties($^) AS start_vertex | GO from $-.id OVER serve REVERSELY WHERE serve._dst != "player100" \
YIELD properties($$).age AS status, $-.sr_start_year AS sr_start_year, serve.start_year AS tr_start_year, [$-.start_vertex, $-.edge0, $-.middle_vertex, properties(edge), properties($$)] AS p, properties($$) AS target | \
GROUP BY $-.status YIELD $-.status AS status, $-.target AS target, collect(p) AS `path`, 'HE-BE10' AS nodeType, max($-.sr_start_year) AS p1, max($-.tr_start_year) AS p2
可以看到聚合之前的query是ok的，我也很蹩脚地窜出来了 path
GO FROM "player100" OVER serve YIELD serve._dst AS id, serve.start_year AS sr_start_year, properties($$) AS middle_vertex, properties(edge) AS edge0, properties($^) AS start_vertex | GO from $-.id OVER serve REVERSELY WHERE serve._dst != "player100" \
YIELD properties($$).age AS status, $-.sr_start_year AS sr_start_year, serve.start_year AS tr_start_year, [$-.start_vertex, $-.edge0, $-.middle_vertex, properties(edge), properties($$)] AS p, properties($$) AS target
+--------+---------------+---------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------+--------------------------------------+
| status | sr_start_year | tr_start_year | p | target |
+--------+---------------+---------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------+--------------------------------------+
| 32 | 1997 | 2018 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2019, start_year: 2018}, {age: 32, name: "Marco Belinelli"}] | {age: 32, name: "Marco Belinelli"} |
| 36 | 1997 | 1999 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2018, start_year: 1999}, {age: 36, name: "Tony Parker"}] | {age: 36, name: "Tony Parker"} |
| 33 | 1997 | 2015 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2019, start_year: 2015}, {age: 33, name: "LaMarcus Aldridge"}] | {age: 33, name: "LaMarcus Aldridge"} |
| 32 | 1997 | 2017 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2019, start_year: 2017}, {age: 32, name: "Rudy Gay"}] | {age: 32, name: "Rudy Gay"} |
| 31 | 1997 | 2010 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2018, start_year: 2010}, {age: 31, name: "Danny Green"}] | {age: 31, name: "Danny Green"} |
| 25 | 1997 | 2014 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2018, start_year: 2014}, {age: 25, name: "Kyle Anderson"}] | {age: 25, name: "Kyle Anderson"} |
| 32 | 1997 | 2013 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2015, start_year: 2013}, {age: 32, name: "Aron Baynes"}] | {age: 32, name: "Aron Baynes"} |
| 36 | 1997 | 2012 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2016, start_year: 2012}, {age: 36, name: "Boris Diaw"}] | {age: 36, name: "Boris Diaw"} |
| 34 | 1997 | 2010 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2015, start_year: 2010}, {age: 34, name: "Tiago Splitter"}] | {age: 34, name: "Tiago Splitter"} |
| 27 | 1997 | 2011 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2015, start_year: 2011}, {age: 27, name: "Cory Joseph"}] | {age: 27, name: "Cory Joseph"} |
| 38 | 1997 | 2015 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2016, start_year: 2015}, {age: 38, name: "David West"}] | {age: 38, name: "David West"} |
| 29 | 1997 | 2015 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2017, start_year: 2015}, {age: 29, name: "Jonathon Simmons"}] | {age: 29, name: "Jonathon Simmons"} |
| 29 | 1997 | 2016 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2019, start_year: 2016}, {age: 29, name: "Dejounte Murray"}] | {age: 29, name: "Dejounte Murray"} |
| 39 | 1997 | 2013 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2013, start_year: 2013}, {age: 39, name: "Tracy McGrady"}] | {age: 39, name: "Tracy McGrady"} |
| 41 | 1997 | 2002 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2018, start_year: 2002}, {age: 41, name: "Manu Ginobili"}] | {age: 41, name: "Manu Ginobili"} |
| 38 | 1997 | 2016 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2019, start_year: 2016}, {age: 38, name: "Paul Gasol"}] | {age: 38, name: "Paul Gasol"} |
| 32 | 1997 | 2013 | [{age: 42, name: "Tim Duncan"}, {end_year: 2016, start_year: 1997}, {name: "Spurs"}, {end_year: 2015, start_year: 2013}, {age: 32, name: "Marco Belinelli"}] | {age: 32, name: "Marco Belinelli"} |
+--------+---------------+---------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------+--------------------------------------+
Got 17 rows (time spent 5882/30980 us)
但是只要下一步开始聚合的时候这里边有两个问题：
所以也只能输出去掉了 path和target 的：
GO FROM "player100" OVER serve YIELD serve._dst AS id, serve.start_year AS sr_start_year, properties($$) AS middle_vertex, properties(edge) AS edge0, properties($^) AS start_vertex | GO from $-.id OVER serve REVERSELY WHERE serve._dst != "player100" \
YIELD properties($$).age AS status, $-.sr_start_year AS sr_start_year, serve.start_year AS tr_start_year, [$-.start_vertex, $-.edge0, $-.middle_vertex, properties(edge), properties($$)] AS p, properties($$) AS target | \
GROUP BY $-.status YIELD $-.status AS status, 'HE-BE10' AS nodeType, max($-.sr_start_year) AS p1, max($-.tr_start_year) AS p2
这个情况下看起来还是用 cypher match 去做。
(root@nebula) [basketballplayer]> GO FROM "player100" OVER serve YIELD serve._dst AS id, serve.start_year AS sr_start_year, properties($$) AS middle_vertex, properties(edge) AS edge0, properties($^) AS start_vertex | GO from $-.id OVER serve REVERSELY WHERE serve._dst != "player100" \
-> YIELD properties($$).age AS status, $-.sr_start_year AS sr_start_year, serve.start_year AS tr_start_year, [$-.start_vertex, $-.edge0, $-.middle_vertex, properties(edge), properties($$)] AS p, properties($$) AS target | \
-> GROUP BY $-.status YIELD $-.status AS status, $-.target AS target, collect(p) AS `path`, 'HE-BE10' AS nodeType, max($-.sr_start_year) AS p1, max($-.tr_start_year) AS p2
[ERROR (-1009)]: SemanticError: Not supported expression `p' for props deduction.
(root@nebula) [basketballplayer]> GO FROM "player100" OVER serve YIELD serve._dst AS id, serve.start_year AS sr_start_year, properties($$) AS middle_vertex, properties(edge) AS edge0, properties($^) AS start_vertex | GO from $-.id OVER serve REVERSELY WHERE serve._dst != "player100" \
-> YIELD properties($$).age AS status, $-.sr_start_year AS sr_start_year, serve.start_year AS tr_start_year, [$-.start_vertex, $-.edge0, $-.middle_vertex, properties(edge), properties($$)] AS p, properties($$) AS target | \
-> GROUP BY $-.status YIELD $-.status AS status, $-.target AS target
[ERROR (-1009)]: SemanticError: Yield non-agg expression `$-.target' must be functionally dependent on items in GROUP BY clause
Thu, 06 Jan 2022 11:32:22 CST
收到，十分感谢。我之前想的把聚合效果拆开试试，类似我上面写的效果，但是发现，需求还有三跳的查询，每一跳的结果都要通过管道传递给下一层，这个很繁琐。
我先用match 试试性能吧，再次感谢。
此话题已在最后回复的 30 天后被自动关闭。不再允许新回复。