The background data quantity is large, the request of data interface, it is slow, when the productio-CodePudding

All the data display on a page, the page like currency charts

CodePudding user response:

Have the paging down paging

CodePudding user response:

If it's the same time a large number of requests, can be set to check code, please input from the user Numbers to examine, judge not robot operation), reduce the same time a large number of requests,

CodePudding user response:

1. Improve the response speed of the API, if the request can be added more load balancing, behind all match several API server,
2. Reduce the json API redundancy to compress the data volume, like your currency graph data should be minutes, don't use the key, the value of the json to store, with a singular set way to store data of a field, such as: time: [1001, 1002, 1003,... , the open: [10.1, 10.2, 10.1,...
3. Calculation of indicators on the front end of charts, do not put in the background, the background only returns based on high and low charge event data, such as
4. Data update, don't again go through the historical data in the next, download the data of the day, only on the front end joining together until you have downloaded the data back,

https://opensource.zealink.com/hqweb/demo/phone.html I do K line graph probably took 4 years of data, you can search, the source at https://github.com/jones2000/HQChart

CodePudding user response:

Refer to the bootstrap paging, simply don't have to write their own

CodePudding user response:

Data is how to apply colours to a drawing of the front desk? Asynchronous page loading can be considered

CodePudding user response:

Load balancing, reduce project cluster pressure

CodePudding user response:

The building Lord how do you solve? I also encountered this problem, return to the data of 300 w +, nearly 70 million, front-end often crash, I really don't know who replied to the paging, reply with no brain, trend chart how paging, paging base, so who can think of, at a minimum, the multidimensional query a graph of a day, unable to load the data in advance, if you want to make pretreatment, the condition of all dimensions, together, so that every query on the data of one day, but the NTH power of 2 dimensions of data save, really doesn't work

CodePudding user response:

Not true data in the Cache

CodePudding user response:

reference LKratos_ eighth floor master elder brother reply:

the building Lord how do you solve? I also encountered this problem, return to the data of 300 w +, nearly 70 million, front-end often crash, I really don't know who replied to the paging, reply with no brain, trend chart how paging, paging base, so who can think of, at a minimum, the multidimensional query a graph of a day, unable to load the data in advance, if you want to make pretreatment, the condition of all dimensions, together, so that every query on the data of one day, but the NTH power of 2 dimensions of data save, really won't work

Being so large, directly generate a json or js files bai,
Or consider splitting, checked by the hour, and then the front stitching data, modify the interactive logic, not to reveal all at a time, if there are multiple lines, can also check every single line, after all, is to see the data can be split,,,

CodePudding user response:

when [b] [/b] [color=# FF9900] [/color]

CodePudding user response:

references to the tenth floor nian_cj response:

Quote: refer to the eighth floor LKratos_ master elder brother reply:

The building Lord how do you solve? I also encountered this problem, return to the data of 300 w +, nearly 70 million, front-end often crash, I really don't know who replied to the paging, reply with no brain, trend chart how paging, paging base, so who can think of, at a minimum, the multidimensional query a graph of a day, unable to load the data in advance, if you want to make pretreatment, the condition of all dimensions, together, so that every query on the data of one day, but the NTH power of 2 dimensions of data save, really won't work

Being so large, directly generate a json or js files bai,
Or consider splitting, checked by the hour, and then the front stitching data, modify the interactive logic, not to reveal all at a time, if there are multiple lines, you can also check every single line, after all, is to see the data can be split, and, [/quote brother, you think about it again, I'm going to see the trend chart of a day, this is the lowest request, 24 hours a day, you save the file, what combination query within 24 hours of time? And you generate all the combination query, the client must need to all of this data? If you don't reveal all? How call graph of a day? Well think deeply
CodePudding user response:
refer to 12th floor LKratos_ master elder brother reply:

Quote: reference to the tenth floor nian_cj response:

Quote: refer to the eighth floor LKratos_ master elder brother reply:

The building Lord how do you solve? I also encountered this problem, return to the data of 300 w +, nearly 70 million, front-end often crash, I really don't know who replied to the paging, reply with no brain, trend chart how paging, paging base, so who can think of, at a minimum, the multidimensional query a graph of a day, unable to load the data in advance, if you want to make pretreatment, the condition of all dimensions, together, so that every query on the data of one day, but the NTH power of 2 dimensions of data save, really won't work

Being so large, directly generate a json or js files bai,
Or consider splitting, checked by the hour, and then the front stitching data, modify the interactive logic, not to reveal all at a time, if there are multiple lines, you can also check every single line, after all, is to see the data can be split,,,

brother, you think about it again, I'm going to see the trend chart of a day, this is the lowest request, 24 hours a day, you save the file, what combination query within 24 hours of time? And you generate all the combination query, the client must need to all of this data? If you don't show all? How is a day of charts? Deep well think

You didn't know what I mean, is not to say that don't show a day, but to make partial load, part load allows users to look at first, and then gradually behind the load, unapt let blank pages for a long time, of course, specific still should see page design and interactive requirements, it is had to choose, as for the documents, in order to make the interface doesn't collapse, after all, the interface to return to such a big data is not very stable, and if you return a file address, and then to download file, interface itself is not a problem, at least not long time hanging in there,
CodePudding user response:
reference 13 floor nian_cj reply:

Quote: refer to 12th floor LKratos_ master elder brother reply:

The
Quote: reference to the tenth floor nian_cj response:

Quote: refer to the eighth floor LKratos_ master elder brother reply:

The building Lord how do you solve? I also encountered this problem, return to the data of 300 w +, nearly 70 million, front-end often crash, I really don't know who replied to the paging, reply with no brain, trend chart how paging, paging base, so who can think of, at a minimum, the multidimensional query a graph of a day, unable to load the data in advance, if you want to make pretreatment, the condition of all dimensions, together, so that every query on the data of one day, but the NTH power of 2 dimensions of data save, really won't work

Being so large, directly generate a json or js files bai,
Or consider splitting, checked by the hour, and then the front stitching data, modify the interactive logic, not to reveal all at a time, if there are multiple lines, you can also check every single line, after all, is to see the data can be split,,,

brother, you think about it again, I'm going to see the trend chart of a day, this is the lowest request, 24 hours a day, you save the file, what combination query within 24 hours of time? And you generate all the combination query, the client must need to all of this data? If you don't show all? How is a day of charts? Deep well think

You didn't know what I mean, is not to say that don't show a day, but to make partial load, part load allows users to look at first, and then gradually behind the load, unapt let blank pages for a long time, of course, specific still should see page design and interactive requirements, it is had to choose, as for the documents, in order to make the interface doesn't collapse, after all, the interface to return to such a big data is not very stable, and if you return a file address, and then to download files, interface itself is not a problem, at least not long time hanging in there,

Before you say the train of thought, I have seen, and the front end echars similar global rendering, the coordinates of watching them into static data, are the coordinates of partial to load this part of the static data... There is nothing wrong with the idea, I'm not very clear our data is dynamic, multi-dimensional real-time query, how partial? An interface is a response to this way of thinking should be no problem, now the interface response for 1 seconds, partial block data, feeling principle is paging, almost hope comment,,, now I do is data cable, but the feeling is not a good solution, if data volume increase, another itself data volume is too big, compressed again still not quick response demand
CodePudding user response:
On the 14th floor LKratos_
reference master elder brother reply:

Quote: refer to the 13th floor nian_cj response:

Quote: refer to 12 floor LKratos_ master elder brother reply:

The
Quote: reference to the tenth floor nian_cj response:

nullnullnullnullnullnullnullnullnullnullnullnullnullnullnullnull