feat(hogql): select statements

frontend/src/queries/schema.json

@@ -1600,34 +1600,8 @@
                     "type": "array"
                 },
                 "response": {
-                    "additionalProperties": false,
-                    "description": "Cached query response",
-                    "properties": {
-                        "columns": {
-                            "items": {
-                                "type": "string"
-                            },
-                            "type": "array"
-                        },
-                        "hasMore": {
-                            "type": "boolean"
-                        },
-                        "results": {
-                            "items": {
-                                "items": {},
-                                "type": "array"
-                            },
-                            "type": "array"
-                        },
-                        "types": {
-                            "items": {
-                                "type": "string"
-                            },
-                            "type": "array"
-                        }
-                    },
-                    "required": ["columns", "types", "results"],
-                    "type": "object"
+                    "$ref": "#/definitions/EventsQueryResponse",
+                    "description": "Cached query response"
                 },
                 "select": {
                     "description": "Return a limited set of data. Required.",
@@ -1647,6 +1621,33 @@
             "required": ["kind", "select"],
             "type": "object"
         },
+        "EventsQueryResponse": {
+            "additionalProperties": false,
+            "properties": {
+                "columns": {
+                    "items": {},
+                    "type": "array"
+                },
+                "hasMore": {
+                    "type": "boolean"
+                },
+                "results": {
+                    "items": {
+                        "items": {},
+                        "type": "array"
+                    },
+                    "type": "array"
+                },
+                "types": {
+                    "items": {
+                        "type": "string"
+                    },
+                    "type": "array"
+                }
+            },
+            "required": ["columns", "types", "results"],
+            "type": "object"
+        },
         "FeaturePropertyFilter": {
             "additionalProperties": false,
             "properties": {

frontend/src/queries/schema.ts

@@ -134,6 +134,13 @@ export interface NewEntityNode extends EntityNode {
     event?: string | null
 }
 
+export interface EventsQueryResponse {
+    columns: any[]
+    types: string[]
+    results: any[][]
+    hasMore?: boolean
+}
+
 export interface EventsQuery extends DataNode {
     kind: NodeKind.EventsQuery
     /** Return a limited set of data. Required. */
@@ -170,12 +177,7 @@ export interface EventsQuery extends DataNode {
     /** Columns to order by */
     orderBy?: string[]
 
-    response?: {
-        columns: string[]
-        types: string[]
-        results: any[][]
-        hasMore?: boolean
-    }
+    response?: EventsQueryResponse
 }
 
 export interface PersonsNode extends DataNode {

posthog/hogql/ast.py

@@ -1,6 +1,6 @@
 import re
 from enum import Enum
-from typing import Any, List, Literal
+from typing import Any, List, Literal, Optional, Union
 
 from pydantic import BaseModel, Extra
 
@@ -102,3 +102,31 @@ class Placeholder(Expr):
 class Call(Expr):
     name: str
     args: List[Expr]
+
+
+class JoinExpr(Expr):
+    table: Optional[Union["SelectQuery", Field]] = None
+    table_final: Optional[bool] = None
+    alias: Optional[str] = None
+    join_type: Optional[str] = None
+    join_constraint: Optional[Expr] = None
+    join_expr: Optional["JoinExpr"] = None
+
+
+class SelectQuery(Expr):
+    select: List[Expr]
+    distinct: Optional[bool] = None
+    select_from: Optional[JoinExpr] = None
+    where: Optional[Expr] = None
+    prewhere: Optional[Expr] = None
+    having: Optional[Expr] = None
+    group_by: Optional[List[Expr]] = None
+    order_by: Optional[List[OrderExpr]] = None
+    limit: Optional[Expr] = None
+    limit_by: Optional[List[Expr]] = None
+    limit_with_ties: Optional[bool] = None
+    offset: Optional[Expr] = None
+
+
+JoinExpr.update_forward_refs(SelectQuery=SelectQuery)
+JoinExpr.update_forward_refs(JoinExpr=JoinExpr)

posthog/hogql/constants.py

@@ -124,3 +124,6 @@
     "person.created_at",
     "person.properties",
 ]
+
+# Never return more rows than this in top level HogQL SELECT statements
+MAX_SELECT_RETURNED_ROWS = 65535

posthog/hogql/context.py

@@ -20,4 +20,9 @@ class HogQLContext:
     field_access_logs: List[HogQLFieldAccess] = field(default_factory=list)
     # Did the last calls to translate_hogql since setting these to False contain any of the following
     found_aggregation: bool = False
+    # Do we need to join the persons table or not
     using_person_on_events: bool = True
+    # If set, allows printing full SELECT queries in ClickHouse
+    select_team_id: Optional[int] = None
+    # Do we apply a limit of MAX_SELECT_RETURNED_ROWS=65535 to the topmost select query?
+    limit_top_select: bool = True

posthog/hogql/grammar/HogQLParser.g4

@@ -21,7 +21,6 @@ selectStmt:
     groupByClause? (WITH (CUBE | ROLLUP))? (WITH TOTALS)?
     havingClause?
     orderByClause?
-    limitByClause?
     limitClause?
     settingsClause?
     ;
@@ -37,8 +36,7 @@ groupByClause: GROUP BY ((CUBE | ROLLUP) LPAREN columnExprList RPAREN | columnEx
 havingClause: HAVING columnExpr;
 orderByClause: ORDER BY orderExprList;
 projectionOrderByClause: ORDER BY columnExprList;
-limitByClause: LIMIT limitExpr BY columnExprList;
-limitClause: LIMIT limitExpr (WITH TIES)?;
+limitClause: LIMIT limitExpr ((WITH TIES) | BY columnExprList)?;
 settingsClause: SETTINGS settingExprList;
 
 joinExpr