Hi, Habr. In the last article, I talked about the initial domain analysis and basic design of our new ECM system. Now I will tell you about the first practical problem that we solved. Namely - about the choice of a way to organize the structure of the storage of business attributes of objects.
, :
. - - - , 1000. , , 15% . .
, , , , .
?
, , . .
, , , , , , , . - , . , , .
, , .
- , - .
. , . , "".
( ) , .
, . "" - . , , , , , . . , . , ?
?
- ( , ) , NoSQL- . JSON , .
, , :
, , . .
- json , .
- , JSON.
:
.
.
, , .
, . , . . .
, , , .
, NoSQL- , . , . Postgre, JSON .
:
:
r_objects, : , .
r_attributes. , . , .
JSON:
objects. , data jsonb .
- .
5 , , .
:
. : , .
. : , . , , .
. : . , ().
20 50 . , , 200 . .
:
() - - . - Postgres , , count - , . 10000- , 10000. , , .
() - .
() - .
- , , , NoSQL - .
, .
40000 :
|
|
|
NoSQL |
|
74 |
66 |
|
970 |
1080 |
, |
42 |
37 |
|
40000 |
40000 |
, , , .
?
, , NoSQL- . , , , . , .