Shulou(Shulou.com)06/02 Report--

This article introduces you how to MONGODB query, the content is very detailed, interested friends can use for reference, I hope it can be helpful to you.

Developers who are used to SQL statements are likely to misunderstand this database when using MONGODB. Correct three points

Although 1 MONGODB is a storage JOSN string, but it is not how you want to save it, although it is more open than the traditional database in dealing with data on the SCHEMA, but if you really open it, the attribution nature of the MONGODB database will put you back to your original form.

Although the query statement of 2 MONGODB is not as good as the traditional SQL statement, you can make so many judgments (just updated, MONGO statement can have IF THEN in the query sentence, contrary to the sky), conversion, but the nature of the database, write MONGODB query statement, also want to have some skills, otherwise the query can let you wait until the sky is dark, this is not to blame MONGODB, it is to blame you do not understand her principle and nature.

3 the use of MONGODB is also standardized, not programmers do what they want, everything in the world has its own rules and attributes, so the use of MONGODB naturally has to be standardized.

The following three aspects from the above to let those who do not know how to use MONGODB, to familiarize themselves with.

ONE and MONGODB have no database SCHEMA. I can save the document of each collection as much as I want. I can save the record of one of my loans in the last docuemnt, and I can save the default record of the repayment in the next Document.

If you ask me, can you do this? from the principle of MONGODB, you can, of course you can, but I want to ask, the database properties of MONGODB, have you forgotten that she is not notepad, she is a data for you to insert that may be dozens of times more than the traditional database, and eventually you have to filter out dozens of pieces of data from these tens of millions, billions of pieces of data, you think so above you. How does each DOCUMENT's unrelated storage method help your system? you might as well use a text file, which is more suitable for you.

MONGODB COLLECTION (table), is the meaning of a table, although a MONGODB table (collection) can make your SCHEMA without any rules, but you should have a spectrum in your heart, what should this collection have, and what role these collection composed of documentS will play in my future data processing.

Let's talk about why there are databases like MONGODB, as well as large amounts of data and micro-services, as well as agile development. Just imagine if you have a project, you don't know how many fields it has, how many fields will be added and deleted in the next year, and you don't know the names of these fields, and you will soon start debugging your project in the current state. And online, how do you feel? do you want to blurt out the English word that starts with F right away?

STOP,MONGODB can help you respond to and deal with such projects quickly, but that doesn't mean MONGODB is a dump and scapegoat for such projects. It is such nonsensical projects and uncertain watch designs that make MONGODB stand out. What, there is no such project, are you sure? )

In addition, the aggregation of information in micro services, if you have such a set of MONGODB to help you communicate, centralize and transparent the information of multiple components in micro services, this is another way to use it, because in micro services, information transmission and information loss already make you feel uncomfortable, and MONGODB, as your housekeeper, saves all this information and makes it easy for you to query with MONGODB sentences. Instead of looking it up in the file log, and the speed is quite fast, us nanosecond speed, hundreds of millions of levels of data, of course, all this is focused on you can understand her structure and use, and even you can use MONGODB as the basic database of BI, and chaotic and complex information will be output through the suite. (MONGO supports BI's own suite)

If English is the common sentence in the world, JOSN is the universal sentence for message transmission in computer programs, and all kinds of messages can pass your data from one microservice world to another through a common JOSN format, regardless of whether it is WINDOWS system, LINUX system, JAVA or .NET CORE or PYTHON, GO, etc., because as long as you conform to the data stored in JOSN format. It can be handled any way in the computer world, because he is Esperanto.

And MONGO DB has a strong ability to deal with JOSN format data, you can use a variety of MONGO DB syntax, JOSN layers of nesting and arrays are naked, you want to present in front of your eyes.

STOP, it goes too far.

The following is how to query the information you want using MONGO DB statements and do so by indexing. So here are two questions: 1 the way of query and 2 the skills of index establishment (MONGODB query is also broad and profound, the index is rich and colorful, the ability is limited, as much as you can)

1 query data, please note that there is no transactionality, that is, there is no multi-table association in MONGODB, your attention only needs to be directed to the table you are looking for.

Note: MONGODB updates quickly, some sentences found on the Internet are out of date or can no longer be used, please check your corresponding version. The following statements are based on MONGODB 3.6as a standard.

To take a simple example, we have another collection with the following data of 2700 million (less data, let's put it together)

We want to find matching data through the filesource and filetype attributes, so we have the following query statement to do this.

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"})

Through the above sentence, the data quickly found out, (voiceover: so simple), of course not, a little bit of difficulty to come, a mouthful of rice.

So the question is, if we query the data in MONGO DB, if we only need to display part of the data, but not all the data, whether we can save the query time.

YES Yes, just like the traditional database query, if you want to display the data, it will be marked, you want to see those, you do not look at, you must not let him SHOW out.

This is MONGODB query rule 1

Use that field, just display that field, don't show it if you don't need it. How to write specific sentences

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"}, {"_ id": 0, "filename": 0, "createdate": 0, "filetype": 0, "filesource": 0})

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"})

What is the difference in the execution efficiency of the statement written in this way? compared here, the efficiency has been increased by 10%, from 0.020 seconds to 0.018 seconds. Of course, if the amount of data is again large and the size of the filtered field is large, the speed will be faster. It can be said that this is no different from the traditional database.

So developers of DEVELOPER, using MONGODB, also do not SHOW the fields you do not want, but prohibit them. The syntax is to add {} at the end and a colon after the field name with 0.

Of course, the developer said, you are not scientific, MONGODB every DOCUEMNT (line), is different, how can I do this?

Good job: then just show the fields you want.

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"}, {"_ id": 1})

、

All you need is to mark the place marked 0 as 1, and as you can see from the screenshot, the execution time has reached 0.08 seconds, which is 60% faster. So you need to know this rule.

So let's go on, make it more difficult, and now we think that the query time is still too long, you (I've always claimed that the query speed of MONGODB can reach nanosecond), let's see, MONGODB query continues to speed up.

Indexes, MONGODB, like traditional databases, have indexes, and there are more types of indexes.

In view of the above query, how do we do it?

In fact, we still need to know

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"}, {"_ id": 0, "filename": 0, "createdate": 0, "filetype": 0, "filesource": 0})

This statement is actually a full table scan, so it is not OK, we need to build an index

(of course, before establishing the index, we should analyze whether the establishment of the index can be more effective, whether the establishment of the index is worth it, today's time is relatively tight, we will not discuss it for the time being, of course, we generally still have to establish the analysis of the sampling rate.) in fact, from the point of view of this sampling rate, adding an index is not ideal. But here we are just a demonstration to build an index.

Db.dbmongolog2.createIndex ({filesource:1,filetype:1}, {name: "idx_dbmongolog2_filesource_filetype", background:true})

It can be verified by the execution plan, and the current query can go to the index that has just been created.

But don't be happy too soon, if you rewrite the query to

Db.dbmongolog2.find ({"filesource": 0, "filetype": "1"}, {"_ id": 0, "filename": 0, "createdate": 0, "filetype": 0, "filesource": 0}) .sort ({_ id:1})

Immediately looking at the implementation plan, no longer take the newly built index, take the OBJECT_ID primary key, of course, the efficiency of execution is worse than the efficiency of full table scanning. What exactly did we do here to stop MONGODB from taking the index we just built, and the answer is sort ({_ id:1})

So in general query statements, do not use sorting if you do not need sorting, otherwise it is very likely that the index you have just worked hard to build will no longer be in effect. But what if it has to be sorted.

After re-establishing the index, our query speed is the same as flying again, see the following figure

Db.dbmongolog2.createIndex ({filesource:1,filetype:1,_id:-1}, {name: "idx_dbmongolog2_filesource_filetype_id", background:true})

The above indexing statement, please note that the number after each field, 1 is ascending,-1 is descending, especially sorting, if often descending, do not build the index into ascending order.

After talking about simple query and index establishment, in fact, a MONGODB database system can be established smoothly, mainly considering your collection, of course, if you only use MONGODB as a log system, you can consider less.

But in fact, MONGODB's collection can do a lot of things, for example, he can easily complete several load JOIN operations on the traditional database in a collection, through nesting, these JOIN will be reflected in a collection, so the query speed will be very fast, which is unmatched by the traditional database. Shortcomings, may lose space, is a way to use space, for time. The data of MONGODB will be compressed automatically and will be compressed as soon as it is entered, so you have to find some shortcomings and say something.

The following is the query and index of the table, and further, since it is a promotion

1 index building skills, we continue to look at the following query statement, we query a time after December 18, 2018 records, and the filetype field, is

Db.dbmongolog2.find ({"createdate": {$gte:ISODate ("2018-12-18T17:39:43.207+08:00")}, "filetype": 1}

Obviously, the query has gone through a full table scan, not an index has been established. Why not? let's write it another way.

Db.dbmongolog2.find ({"filetype": 1, "createdate": {$gte:ISODate ("2018-12-18T17:39:43.207+08:00")}}, {"_ id": 1}) .explain ()

In another way, check the execution plan, no, or full table scan.

Here is a summary of MONGODB's law of indexing.

Index establishment skills, you can first calculate the sampling rate, distribution of discrete fields in front, in line with the field index, must-check fields in front of a query can be multiple fields, multiple indexes, you can use the intersection of indexes (the concept of INDEX MERGE in MYSQL is a bit like), how to use depends on the choice of optimization engine. But if a composite index can solve it, it's best not to use it.

Index Intersection .

So a MONGODB looks like a document database, and the overall execution file is only 100 megabytes, but the contents are no less knowledgeable than ORACLE SQL SERVER and are enough to drink a pot.

In addition, a small MONGODB can also do aggregation, similar to traditional database GROUP BY HAVING SUM, AVAGE and other operations, so this MONGODB is by no means good. Since it starts to support transactions, NO-SQL database will not only grab the non-relational market, but also take a share of your relational market.

Aggregation operation of MONGODB

Db.dbmongolog2.aggregate ([{"$match": {"createdate": {$gte:ISODate ("2018-12-18T17:39:43.207+08:00")}}, {$group: {_ id: {filetype: "$filetype", filesource: "$filesource"}, "count": {"$sum": 1}}]. Explain ()

On how to conduct MONGODB query to share here, I hope that the above content can be of some help to you, can learn more knowledge. If you think the article is good, you can share it for more people to see.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.