• I'm just assuming you should actually read https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-and-highlighting-code-blocks first. Your code blocks aren't displaying properly.

how to use SHARDING ALGORITHM Type: INTERVAL two database？

      app_user:
        actualDataNodes: <INTERVAL>P=ds0.app_user_;SP=yyyyMM;DIA=1;DIU=Months;DL=202401;DU=202412 
        tableStrategy:
          standard:
            shardingColumn: updatetime
            shardingAlgorithmName: app_user_table_inline
        keyGenerateStrategy:
          column: id
          keyGeneratorName: snowflake

it work，but i have 2 database,I don't know how to configure it.
I have tried the following configurations, but none of them work:

<INTERVAL>P=ds0.app_user_;SP=yyyyMM;DIA=1;DIU=Months;DL=202401;DU=202412,<INTERVAL>P=ds1.app_user_;SP=yyyyMM;DIA=1;DIU=Months;DL=202401;DU=202412

<INTERVAL>P=ds${0..1}.app_user_;SP=yyyyMM;DIA=1;DIU=Months;DL=202401;DU=202412 

Canonly be used for a single database?

• I'm actually the original author of the Row-Value expression implementation of INTERVAL. The relevant SPI has nothing to do with the sharding algorithm, which is the SPI for Row-Value expression. See https://shardingsphere.apache.org/document/current/cn/user-manual/common-config/builtin-algorithm/expr/ .
• The logic of INTERVAL is located in org.apache.shardingsphere.infra.expr.interval.IntervalInlineExpressionParser, which is actually an extraction of the logic in an overly complex unit test, and that unit test only involves a single database. You can get a general understanding of what this unit test is from https://www.yuque.com/linghengqian/meve2v/cgi5en .
• I would say that this SPI implementation class was indeed not originally designed with multiple databases in mind.
• But then again, you can always create your own SPI implementation class of org.apache.shardingsphere.infra.expr.spi.InlineExpressionParser to generate a java.util.List<String> from some more complex business logic, including connecting to a redis server to get a java.util.List<String>.
• LITERAL, INTERVAL and GROOVY are always available. I personally do not recommend using ESPRESSO at the current point in time, as ESPRESSO is involved in multiple issues with the GraalVM community.

Thank you very much!

• This is indeed a documentation issue, I opened Add documentation for creating SPI implementations of InlineExpressionParser #33653 .