RocMarshal commented on a change in pull request #12798:
URL: https://github.com/apache/flink/pull/12798#discussion_r450299225
##########
File path: docs/dev/table/streaming/match_recognize.zh.md
##########
@@ -1091,42 +983,31 @@ DEFINE
C as C.price > 20
{% endhighlight %}
-<span class="label label-danger">Attention</span> Please note that the
`MATCH_RECOGNIZE` clause
-does not use a configured [state retention
time](query_configuration.html#idle-state-retention-time).
-One may want to use the `WITHIN` [clause](#time-constraint) for this purpose.
+<span class="label label-danger">注意</span> 请注意，`MATCH_RECOGNIZE` 子句未使用配置的
[state retention time]({% link dev/table/streaming/query_configuration.zh.md
%}#idle-state-retention-time)。为此，可能需要使用 `WITHIN` [clause](#known-limitations)。
+
+<a name="known-limitations"></a>
-Known Limitations
+已知的局限
-----------------
-Flink's implementation of the `MATCH_RECOGNIZE` clause is an ongoing effort,
and some features of
-the SQL standard are not yet supported.
-
-Unsupported features include:
-* Pattern expressions:
- * Pattern groups - this means that e.g. quantifiers can not be applied to a
subsequence of the
- pattern. Thus, `(A (B C)+)` is not a valid pattern.
- * Alterations - patterns like `PATTERN((A B | C D) E)`, which means that
either a subsequence
- `A B` or `C D` has to be found before looking for the `E` row.
- * `PERMUTE` operator - which is equivalent to all permutations of variables
that it was applied
- to e.g. `PATTERN (PERMUTE (A, B, C))` = `PATTERN (A B C | A C B | B A C |
B C A | C A B | C B A)`.
- * Anchors - `^, $`, which denote beginning/end of a partition, those do not
make sense in the
- streaming context and will not be supported.
- * Exclusion - `PATTERN ({- A -} B)` meaning that `A` will be looked for but
will not participate
- in the output. This works only for the `ALL ROWS PER MATCH` mode.
- * Reluctant optional quantifier - `PATTERN A??` only the greedy optional
quantifier is supported.
-* `ALL ROWS PER MATCH` output mode - which produces an output row for every
row that participated
- in the creation of a found match. This also means:
- * that the only supported semantic for the `MEASURES` clause is `FINAL`
- * `CLASSIFIER` function, which returns the pattern variable that a row was
mapped to, is not yet
- supported.
-* `SUBSET` - which allows creating logical groups of pattern variables and
using those groups in
- the `DEFINE` and `MEASURES` clauses.
-* Physical offsets - `PREV/NEXT`, which indexes all events seen rather than
only those that were
- mapped to a pattern variable (as in [logical offsets](#logical-offsets)
case).
-* Extracting time attributes - there is currently no possibility to get a time
attribute for
- subsequent time-based operations.
-* `MATCH_RECOGNIZE` is supported only for SQL. There is no equivalent in the
Table API.
+Flink 对 `MATCH_RECOGNIZE` 子句的实现是一项持续的工作，目前尚不支持SQL标准的某些功能。
Review comment:
```suggestion
Flink 对 `MATCH_RECOGNIZE` 子句的实现是一项持续的工作，目前尚不支持 SQL 标准的某些功能。
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
