This document will introduce the naming specifications and some limitations for creating and changing databases and tables in Tencent Cloud TCHouse-D.
Naming Specifications
Database
Name: The table name must start with a letter or an underscore. It can contain letters, numbers, and underscores, with a length of 1 to 64 characters.
Description: It can contain up to 2,048 characters.
Limitations: The same data link cannot have identical database names.
Data Table/View
Name: The table name must start with a letter or an underscore. It can contain letters, numbers, and underscores, with a length of 1 to 64 characters.
Description: It can contain up to 2,048 characters.
Limitations: The same database cannot have identical data table names.
Attribute Column
Name: The column name must start with a letter or an underscore. It can contain letters, numbers, and underscores, with a length of 1 to 256 characters.
Description: It can contain up to 256 characters.
Limitations: The same data table cannot have identical data column names.
Partition
Partition feld name: Length should be 1 to 256 characters.
Usage recommendations:
1. When the single table data volume is below 200 million records, for convenience, you can choose not to set partitions and use bucketing instead;
2. When the partition field has a continuous data range (such as date and ID), it is recommended to choose Range partition;
3. When the data has discrete values (such as country, region, and status), it is recommended to choose LIST partition.
Bucketing
Bucket field name: Length should be 1 to 256 characters
Usage recommendations:
1. The bucket key can be one or multiple; multiple keys ensure a more balanced data distribution, while a single key is easier to match.
2. Selection of bucket columns: generally choose fields with high differentiation/cardinality and even hash to avoid data skew; they should also be frequently used fields to improve query efficiency;
3. The number of buckets created should not be too many or too few; it is recommended to keep each bucket between 1-10 GB.
Use Limits
Quantity Limit of Database/Table/Column/Partition
|
Number of catalogs that can be created | 20 |
Number of databases per cluster | 10,000 |
Number of databases per user | 1,000 |
Number of data tables per cluster | 100,000 |
Number of data tables per user | 10,000 |
Number of data tables per database | 4,096 |
Number of fields per table | 2,048 |
Number of partitions per table | 10,000 |
Number of views per cluster | 10,000 |
Maximum view nesting level | 10 |
Number of materialized views per cluster | 10,000 |
Query Limits
|
Daily data query volume | No limit | - |
Number of tables in a single SQL query | 32 | - |
Single SQL query timeout | 900 seconds | - |
Maximum number of files per external table | No limit | - |
DML/DML Limits
|
Simultaneous schema changes per cluster | 20 | - |
Simultaneous schema changes per table | 1 | - |
Data write concurrency per table (per second) | 10 | High-frequency writes are not recommended; it is advised to batch write appropriately: the total number of imports per table should not exceed 20 times per minute. |
Data write concurrency for the same key per table (per second) | 1 | The same key in one table cannot be updated simultaneously; otherwise, it may cause data inaccuracy. |