如何解决CouchDB 过滤器和带有视图和列表的分页
我们正在尝试使用 CouchDB 和 Golang 在 HL Fabric 中构建一个应用程序,它已经增长了太多,现在我们有很多数据类型,我们需要在数据库中进行连接等价物来过滤和显示数据.
我们的尝试:
使用 Mango Queries 通过代码加入链码(这似乎不是一个好的解决方案,但它是唯一真正有效的解决方案) 这是,获取父级,循环获取子级的结果(如果只是列出)
在过滤的情况下,事情变得更加复杂,因为我们需要按组过滤子项,获取 parentID,在另一个查询中使用那些带有我们想要避免的许多 ID 的可怕的“$or”
通过视图和列表加入 CouchDB
所以,我们提出了这个“解决方案”,但它有很多问题
// Collaborator document
{
"_id": "john.doe@email.com","_rev": "33-0a7f4e6f6c4a60d6ffded693ed794cb6","docType": "collaborator","id": "john.doe@email.com","levelID": "1@level",// this is what we want to filter events by (1@level,2@level,...)
"name": "John","surname": "Doe","~version": "CgQCAsIA"
}
// Event document
{
"_id": "0160ca22-6ffb-40d8-a8ad-e13ce5f0a0ef","_rev": "1-a23a010bcadf8b7969fb7a5aba93456a","creation": "2021-04-14T14:29:34-0300","docType": "event","id": "0160ca22-6ffb-40d8-a8ad-e13ce5f0a0ef","movement": "transfer",// we want to filter events by this field also
"from": "jane.doe@email.com","to": "john.doe@email.com",// link to collaborator document
"quantity": 10,"~version": "CgQCAk4A"
}
// Our view function
function (doc) {
if (doc.docType == "event"){
emit([doc.creation,doc.id],{_id: doc.id});
emit([doc.creation,{_id: doc.to});
}
}
// Our list function
function (head,req) {
var row,last_key;
var event,to;
var key;
var query = {};
query.movement = req.query.movement;
query.levelID = req.query.levelID;
send('{"records" : [');
while ((row = getRow())) {
key = row.key[row.key.length-1];
if (last_key != key) {
if (last_key != "undefined") {
sendJSON();
event = undefined;
to = undefined;
}
last_key = key;
}
if (row.doc) {
if (row.doc.docType == "event") {
event = row.doc;
} else if (row.doc.docType == "collaborator"){
to = row.doc;
}
}
}
sendJSON();
send(toJSON({}));
send("]}");
function sendJSON() {
var pass = true;
if (event && to) {
if (query.movement && event.movement != query.movement) {
pass = false;
}
if (query.levelID && to.levelID != query.levelID) {
pass = false;
}
if (pass) {
event.to = to;
}
} else {
pass = false;
}
if (pass) {
send(toJSON(event));
send(",");
}
}
};
// Then we call it with
http://127.0.0.1:5984/testchannel_huenei/_design/new_design/_list/result/events
?include_docs=true&levelID=1@level&movement=transfer
&startkey=["3000-01-01T24:00:00-0300",{}]&endkey=["1999-01-01T00:00:00-0300"]&descending=true
// And the result is something like this
{
"records": [{
"_id": "0160ca22-6ffb-40d8-a8ad-e13ce5f0a0ef","movement": "transfer"
"from": "jane.doe@email.com","to": {
"_id": "john.doe@email.com","name": "John","~version": "CgQCAsIA"
},"quantity": 10,"~version": "CgQCAk4A"
},{}
]
}
它可以工作,但看起来太复杂了,当使用 SQL 时,它可以通过简单的连接/查询来解决。深度嵌套的文档呢?
我们知道可以在 url 中过滤“移动”,但实际上我们还有许多其他字段要过滤,正如我们所读到的,键必须是有序的。
这意味着,如果我有emit([doc.to,doc.movement,null),我需要先指定“to”,即&startkey=[???,"transfer"]&endkey=[???,"transfer",{}]
我们尝试使用 &startkey=[null,"transfer"]&endkey=[\ufff0,{}]
,但没有用
但无论如何,嵌套文档的过滤是最有问题的
另一个主要问题是我们无法对结果进行分页,因为关键字“limit”仅在视图中有效,而在列表中无效
人们如何在真实的 Hyperledger Fabric 世界场景中解决这些问题?或者甚至在一个普通的 CouchDB 场景中?
谢谢
版权声明:本文内容由互联网用户自发贡献,该文观点与技术仅代表作者本人。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容, 请发送邮件至 dio@foxmail.com 举报,一经查实,本站将立刻删除。