# Copyright (c) 2022, Frappe Technologies Pvt. Ltd. and Contributors # License: MIT. See LICENSE import itertools import json import random import re import string import traceback import warnings from collections.abc import Iterable, Sequence from contextlib import contextmanager, suppress from time import time from typing import TYPE_CHECKING, Any from pypika.queries import QueryBuilder import frappe import frappe.defaults from frappe import _, _dict from frappe.database.utils import ( DefaultOrderBy, EmptyQueryValues, FallBackDateTimeStr, FilterValue, LazyMogrify, Query, QueryValues, convert_to_value, get_query_type, is_query_type, ) from frappe.exceptions import DoesNotExistError, ImplicitCommitError from frappe.monitor import get_trace_id from frappe.query_builder import Case from frappe.query_builder.functions import Count from frappe.utils import ( CallbackManager, cint, get_datetime, get_table_name, getdate, now, recursive_defaultdict, sbool, ) from frappe.utils import cast as cast_fieldtype if TYPE_CHECKING: from sqlite3 import Connection as SQLiteConnection from sqlite3 import Cursor as SQLiteCursor from MySQLdb.connections import Connection as MySQLdbConnection from MySQLdb.cursors import Cursor as MySQLdbCursor from psycopg2 import connection as PostgresConnection from psycopg2 import cursor as PostgresCursor from pymysql.connections import Connection as MariadbConnection from pymysql.cursors import Cursor as MariadbCursor ALLOWED_TYPES_FOR_VALUES = tuple | list | dict IFNULL_PATTERN = re.compile(r"ifnull\(", flags=re.IGNORECASE) INDEX_PATTERN = re.compile(r"\s*\([^)]+\)\s*") SINGLE_WORD_PATTERN = re.compile(r'([`"]?)(tab([A-Z]\w+))\1') MULTI_WORD_PATTERN = re.compile(r'([`"])(tab([A-Z]\w+)( [A-Z]\w+)+)\1') # Query Types DDL_QUERY_TYPES = frozenset(("alter", "drop", "create", "truncate", "rename")) IMPLICIT_COMMIT_QUERY_TYPES = frozenset(("start", "alter", "drop", "create", "begin", "truncate")) CREATE_OR_DROP = frozenset(("create", "drop")) COMMIT_OR_ROLLBACK = frozenset(("commit", "rollback")) WRITE_QUERY_TYPES = frozenset(("update", "insert", "delete")) QUERY_TYPES_FOR_LOG_TOUCHED_TABLES = frozenset(("insert", "delete", "update", "alter", "drop", "rename")) SQL_ITERATOR_BATCH_SIZE = 1000 TRANSACTION_DISABLED_MSG = """Commit/rollback are disabled during certain events. This command will be ignored. Commit/Rollback from here WILL CAUSE very hard to debug problems with atomicity and concurrent data update bugs.""" class Database: """ Open a database connection with the given parmeters, if use_default is True, use the login details from `conf.py`. This is called by the request handler and is accessible using the `db` global variable. the `sql` method is also global to run queries """ VARCHAR_LEN = 140 MAX_COLUMN_LENGTH = 64 OPTIONAL_COLUMNS = ("_user_tags", "_comments", "_assign", "_liked_by") DEFAULT_SHORTCUTS = ("_Login", "__user", "_Full Name", "Today", "__today", "now", "Now") STANDARD_VARCHAR_COLUMNS = ("name", "owner", "modified_by") DEFAULT_COLUMNS = ("name", "creation", "modified", "modified_by", "owner", "docstatus", "idx") CHILD_TABLE_COLUMNS = ("parent", "parenttype", "parentfield") MAX_WRITES_PER_TRANSACTION = 200_000 class InvalidColumnName(frappe.ValidationError): pass def __init__( self, socket=None, host=None, user=None, password=None, port=None, cur_db_name=None, ): self.setup_type_map() self.socket = socket self.host = host self.port = port self.user = user self.password = password self.cur_db_name = cur_db_name self._conn = None self.transaction_writes = 0 self.auto_commit_on_many_writes = 0 self.value_cache = recursive_defaultdict() self.logger = frappe.logger("database") self.logger.setLevel("WARNING") self.before_commit = CallbackManager() self.after_commit = CallbackManager() self.before_rollback = CallbackManager() self.after_rollback = CallbackManager() # Setting this to true will disable full rollback and commit # You can still use savepoint with partial rollback. self._disable_transaction_control = 0 # self.db_type: str # self.last_query (lazy) attribute of last sql query executed def setup_type_map(self): pass def connect(self): """Connects to a database as set in `site_config.json`.""" self._conn: MySQLdbConnection | MariadbConnection | PostgresConnection | SQLiteConnection = ( self.get_connection() ) self._cursor: MySQLdbCursor | MariadbCursor | PostgresCursor | SQLiteCursor = self._conn.cursor() try: if execution_timeout := get_query_execution_timeout(): self.set_execution_timeout(execution_timeout) except Exception as e: self.logger.warning(f"Couldn't set execution timeout {e}") def set_execution_timeout(self, seconds: int): """Set session speicifc timeout on exeuction of statements. If any statement takes more time it will be killed along with entire transaction.""" raise NotImplementedError def use(self, db_name): """`USE` db_name.""" self._conn.select_db(db_name) self.cur_db_name = db_name def get_connection(self): """Return a Database connection object that conforms with https://peps.python.org/pep-0249/#connection-objects.""" raise NotImplementedError def get_database_size(self): raise NotImplementedError def _transform_query(self, query: Query, values: QueryValues) -> tuple: return query, values def _transform_result(self, result: list[tuple] | tuple[tuple]) -> tuple[tuple]: return result def _clean_up(self): pass def sql( self, query: Query, values: QueryValues = EmptyQueryValues, *, as_dict=0, as_list=0, debug=0, ignore_ddl=0, auto_commit=0, update=None, explain=False, run=True, pluck=False, as_iterator=False, ): """Execute a SQL query and fetch all rows. :param query: SQL query. :param values: Tuple / List / Dict of values to be escaped and substituted in the query. :param as_dict: Return as a dictionary. :param as_list: Always return as a list. :param debug: Print query and `EXPLAIN` in debug log. :param ignore_ddl: Catch exception if table, column missing. :param auto_commit: Commit after executing the query. :param update: Update this dict to all rows (if returned `as_dict`). :param run: Return query without executing it if False. :param pluck: Get the plucked field only. :param explain: Print `EXPLAIN` in error log. :param as_iterator: Returns iterator over results instead of fetching all results at once. This should be used with unbuffered cursor as default cursors used by pymysql and postgres buffer the results internally. See `Database.unbuffered_cursor`. Examples: # return customer names as dicts frappe.db.sql("select name from tabCustomer", as_dict=True) # return names beginning with a frappe.db.sql("select name from tabCustomer where name like %s", "a%") # values as dict frappe.db.sql("select name from tabCustomer where name like %(name)s and owner=%(owner)s", {"name": "a%", "owner":"test@example.com"}) """ if isinstance(query, QueryBuilder): frappe.log("Use run method to execute SQL queries generated by Query Builder") debug = debug or getattr(self, "debug", False) query = str(query) if not run: return query query_type = get_query_type(query) if explain: if debug and query_type == "select": self.explain_query(query, values) return # remove whitespace / indentation from start and end of query # and replace ifnull in query with coalesce query = IFNULL_PATTERN.sub("coalesce(", query.strip()) if not self._conn: self.connect() # in transaction validations self.check_transaction_status(query, query_type) self.clear_db_table_cache(query_type) if auto_commit: self.commit() if debug: time_start = time() if values == EmptyQueryValues: values = None elif not isinstance(values, ALLOWED_TYPES_FOR_VALUES): values = (values,) query, values = self._transform_query(query, values) if trace_id := get_trace_id(): query += f" /* FRAPPE_TRACE_ID: {trace_id} */" try: self.execute_query(query, values) except Exception as e: if self.is_syntax_error(e): frappe.log(f"Syntax error in query:\n{query} {values or ''}") elif self.is_deadlocked(e): raise frappe.QueryDeadlockError(e) from e elif self.is_timedout(e): raise frappe.QueryTimeoutError(e) from e elif self.is_read_only_mode_error(e): frappe.throw( _( "Site is running in read only mode for maintenance or site update, this action can not be performed right now. Please try again later." ), title=_("In Read Only Mode"), exc=frappe.InReadOnlyMode, ) # TODO: added temporarily elif self.db_type == "postgres": traceback.print_stack() frappe.log(f"Error in query:\n{e}") raise elif isinstance(e, self.ProgrammingError): if frappe.conf.developer_mode: traceback.print_stack() frappe.log(f"Error in query:\n{query, values}") raise if not ( ignore_ddl and (self.is_missing_column(e) or self.is_table_missing(e) or self.cant_drop_field_or_key(e)) ): raise if debug: time_end = time() frappe.log(f"Execution time: {(time_end - time_start) * 1000:.3f} ms") self.log_query(query, query_type, values, debug) if auto_commit: self.commit() if not self._cursor.description: return () if as_iterator: return self._return_as_iterator(pluck=pluck, as_dict=as_dict, as_list=as_list, update=update) last_result = self._transform_result(self._cursor.fetchall()) if pluck: last_result = [r[0] for r in last_result] self._clean_up() return last_result # scrub output if required if as_dict: last_result = self.fetch_as_dict(last_result) if update: for r in last_result: r.update(update) elif as_list: last_result = self.convert_to_lists(last_result) self._clean_up() return last_result def _return_as_iterator(self, *, pluck, as_dict, as_list, update): while result := self._transform_result(self._cursor.fetchmany(SQL_ITERATOR_BATCH_SIZE)): if pluck: for row in result: yield row[0] elif as_dict: keys = [column[0] for column in self._cursor.description] for row in result: row = _dict(zip(keys, row, strict=False)) if update: row.update(update) yield row elif as_list: for row in result: yield list(row) else: frappe.throw(_("`as_iterator` only works with `as_list=True` or `as_dict=True`")) self._clean_up() def execute_query(self, query, values=None): return self._cursor.execute(query, values) def _log_query( self, mogrified_query: str, query_type: str, debug: bool = False, unmogrified_query: str = "", ) -> None: """Takes the query and logs it to various interfaces according to the settings.""" _query = None conf = frappe.local.conf if conf.allow_tests and get_print_sql_flag(): _query = _query or str(mogrified_query) print(_query) if debug: _query = _query or str(mogrified_query) frappe.log(_query) if conf.logging == 2: _query = _query or str(mogrified_query) frappe.log(f"#### query\n{_query}\n####") if query_type in DDL_QUERY_TYPES: _query = _query or str(mogrified_query) self.logger.warning("DDL Query made to DB:\n" + _query) if frappe.local.flags.in_migrate: _query = _query or str(mogrified_query) self.log_touched_tables(_query, query_type) def log_query(self, query: str, query_type: str, values: QueryValues = None, debug: bool = False) -> str: mogrified_query = self.lazy_mogrify(query, values) self._log_query(mogrified_query, query_type, debug, query) return mogrified_query def mogrify(self, query: Query, values: QueryValues): """build the query string with values""" if not values: return query try: return self._cursor.mogrify(query, values) except AttributeError: if isinstance(values, dict): return query % { k: frappe.db.escape(v) if isinstance(v, str) else v for k, v in values.items() } elif isinstance(values, list | tuple): return query % tuple(frappe.db.escape(v) if isinstance(v, str) else v for v in values) return query, values def lazy_mogrify(self, query: Query, values: QueryValues) -> LazyMogrify: """Wrap the object with str to generate mogrified query.""" return LazyMogrify(query, values) def explain_query(self, query, values=EmptyQueryValues): """Print `EXPLAIN` in error log.""" frappe.log("--- query explain ---") try: results = self.sql(f"EXPLAIN {query}", values, as_dict=1) except Exception as e: frappe.log(f"error in query explain: {e}") else: frappe.log(json.dumps(results, indent=1)) frappe.log("--- query explain end ---") def sql_list(self, query, values=(), debug=False, **kwargs): """Return data as list of single elements (first column). Example: # doctypes = ["DocType", "DocField", "User", ...] doctypes = frappe.db.sql_list("select name from DocType") """ return self.sql(query, values, **kwargs, debug=debug, pluck=True) def sql_ddl(self, query, debug=False): """Commit and execute a query. DDL (Data Definition Language) queries that alter schema autocommit in MariaDB.""" transaction_control = self._disable_transaction_control self._disable_transaction_control = 0 self.commit() self.sql(query, debug=debug) self._disable_transaction_control = transaction_control def check_transaction_status(self, query: str, query_type: str | None = None): """Raises exception if more than 200,000 `INSERT`, `UPDATE` queries are executed in one transaction. This is to ensure that writes are always flushed otherwise this could cause the system to hang.""" if not query_type: query_type = get_query_type(query) self.check_implicit_commit(query, query_type) if query_type in COMMIT_OR_ROLLBACK: self.transaction_writes = 0 return if query_type in WRITE_QUERY_TYPES: self.transaction_writes += 1 if self.transaction_writes > self.MAX_WRITES_PER_TRANSACTION: if self.auto_commit_on_many_writes: self.commit() else: msg = "

" + _("Too many changes to database in single action.") + "
" msg += _("The changes have been reverted.") + "
" raise frappe.TooManyWritesError(msg) def check_implicit_commit(self, query: str, query_type: str): if query_type in IMPLICIT_COMMIT_QUERY_TYPES and self.transaction_writes: raise ImplicitCommitError("This statement can cause implicit commit", query) def fetch_as_dict(self, result) -> list[_dict]: """Internal. Convert results to dict.""" if not result: return [] keys = [column[0] for column in self._cursor.description] return [_dict(zip(keys, row, strict=False)) for row in result] @staticmethod def clear_db_table_cache(query_type: str): if query_type in CREATE_OR_DROP: frappe.client_cache.delete_value("db_tables") def get_description(self): """Return result metadata.""" return self._cursor.description @staticmethod def convert_to_lists(res): """Convert tuple output to lists (internal).""" return [[value for value in row] for row in res] def get(self, doctype, filters=None, as_dict=True, cache=False): """Return `get_value` with fieldname='*'.""" return self.get_value(doctype, filters, "*", as_dict=as_dict, cache=cache) def get_value( self, doctype: str, filters: FilterValue | dict | list | None = None, fieldname: str | list[str] = "name", ignore: bool = False, as_dict: bool = False, debug: bool = False, order_by: str = DefaultOrderBy, cache: bool = False, for_update: bool = False, *, run: bool = True, pluck: bool = False, distinct: bool = False, skip_locked: bool = False, wait: bool = True, ): """Return a document property or list of properties. :param doctype: DocType name. :param filters: Filters like `{"x":"y"}` or name of the document. `None` if Single DocType. :param fieldname: Column name. :param ignore: Don't raise exception if table, column is missing. :param as_dict: Return values as dict. :param debug: Print query in error log. :param order_by: Column to order by :param cache: Use cached results fetched during current job/request :param pluck: pluck first column instead of returning as nested list or dict. :param for_update: All the affected/read rows will be locked. :param skip_locked: Skip selecting currently locked rows. :param wait: Wait for aquiring lock Example: # return first customer starting with a frappe.db.get_value("Customer", {"name": ("like a%")}) # return last login of **User** `test@example.com` frappe.db.get_value("User", "test@example.com", "last_login") last_login, last_ip = frappe.db.get_value("User", "test@example.com", ["last_login", "last_ip"]) # returns default date_format frappe.db.get_value("System Settings", None, "date_format") """ result = self.get_values( doctype, filters, fieldname, ignore, as_dict, debug, order_by, cache=cache, for_update=for_update, run=run, pluck=pluck, distinct=distinct, limit=1, skip_locked=skip_locked, wait=wait, ) if not run: return result if not result: return None row = result[0] if as_dict or len(row) > 1: return row # single field is requested, send it without wrapping in containers return row[0] def get_values( self, doctype: str, filters: FilterValue | dict | list | None = None, fieldname: str | list[str] = "name", ignore: bool = False, as_dict: bool = False, debug: bool = False, order_by: str = DefaultOrderBy, update: dict | None = None, cache: bool = False, for_update: bool = False, *, run: bool = True, pluck: bool = False, distinct: bool = False, limit: int | None = None, skip_locked: bool = False, wait: bool = True, ): """Return multiple document properties. :param doctype: DocType name. :param filters: Filters like `{"x":"y"}` or name of the document. :param fieldname: Column name. :param ignore: Don't raise exception if table, column is missing. :param as_dict: Return values as dict. :param debug: Print query in error log. :param order_by: Column to order by, :param distinct: Get Distinct results. Example: # return first customer starting with a customers = frappe.db.get_values("Customer", {"name": ("like a%")}) # return last login of **User** `test@example.com` user = frappe.db.get_values("User", "test@example.com", "*")[0] """ out = None if cache and isinstance(filters, str) and fieldname in self.value_cache[doctype][filters]: return self.value_cache[doctype][filters][fieldname] if distinct: order_by = None if isinstance(filters, list): if filters := list(f for f in filters if f is not None): out = frappe.qb.get_query( table=doctype, fields=fieldname, filters=filters, order_by=order_by, distinct=distinct, limit=limit, validate_filters=True, for_update=for_update, skip_locked=skip_locked, wait=True, ).run(debug=debug, run=run, as_dict=as_dict, pluck=pluck) else: out = {} else: if (filters is not None) and (filters != doctype or doctype == "DocType"): try: if order_by: order_by = "creation" if order_by == DefaultOrderBy else order_by query = frappe.qb.get_query( table=doctype, filters=filters, order_by=order_by, for_update=for_update, skip_locked=skip_locked, wait=wait, fields=fieldname, distinct=distinct, limit=limit, validate_filters=True, ) if isinstance(fieldname, str) and fieldname == "*": as_dict = True out = query.run(as_dict=as_dict, debug=debug, update=update, run=run, pluck=pluck) except Exception as e: if ignore and ( frappe.db.is_missing_column(e) or frappe.db.is_table_missing(e) or str(e).startswith("Invalid DocType") ): out = None elif (not ignore) and frappe.db.is_table_missing(e): # table not found, look in singles fields = ( [fieldname] if (isinstance(fieldname, str) and fieldname != "*") else fieldname ) out = self.get_values_from_single( fields, filters, doctype, as_dict, debug, update, run=run, distinct=distinct, ) else: raise else: fields = [fieldname] if (isinstance(fieldname, str) and fieldname != "*") else fieldname out = self.get_values_from_single( fields, filters, doctype, as_dict, debug, update, run=run, pluck=pluck, distinct=distinct, ) if cache and isinstance(filters, str): self.value_cache[doctype][filters][fieldname] = out return out def get_values_from_single( self, fields, filters, doctype, as_dict=False, debug=False, update=None, *, run=True, pluck=False, distinct=False, ): """Get values from `tabSingles` (Single DocTypes) (internal). :param fields: List of fields, :param filters: Filters (dict). :param doctype: DocType name. """ if fields == "*" or isinstance(filters, dict): # check if single doc matches with filters values = self.get_singles_dict(doctype) if isinstance(filters, dict): for key, value in filters.items(): if values.get(key) != value: return [] if as_dict: return [values] if values else [] if isinstance(fields, list): return [list(map(values.get, fields))] else: r = frappe.qb.get_query( "Singles", filters={"field": ("in", tuple(fields)), "doctype": doctype}, fields=["field", "value"], distinct=distinct, ).run(pluck=pluck, debug=debug, as_dict=False) if not run: return r if not r: return [] r = _dict(r) if update: r.update(update) if not as_dict: return [[r.get(field) for field in fields]] return [r] def get_singles_dict(self, doctype, debug=False, *, for_update=False, cast=False): """Get Single DocType as dict. :param doctype: DocType of the single object whose value is requested :param debug: Execute query in debug mode - print to STDOUT :param for_update: Take `FOR UPDATE` lock on the records :param cast: Cast values to Python data types based on field type Example: # Get coulmn and value of the single doctype Accounts Settings account_settings = frappe.db.get_singles_dict("Accounts Settings") """ queried_result = frappe.qb.get_query( "Singles", filters={"doctype": doctype}, fields=["field", "value"], for_update=for_update, ).run(debug=debug) if not cast: return _dict(queried_result) try: meta = frappe.get_meta(doctype) except DoesNotExistError: return _dict(queried_result) return_value = _dict() for fieldname, value in queried_result: if df := meta.get_field(fieldname): casted_value = cast_fieldtype(df.fieldtype, value) else: casted_value = value return_value[fieldname] = casted_value return return_value @staticmethod def get_all(*args, **kwargs): return frappe.get_all(*args, **kwargs) @staticmethod def get_list(*args, **kwargs): return frappe.get_list(*args, **kwargs) @staticmethod def _get_update_dict( fieldname: str | dict, value: Any, *, modified: str, modified_by: str, update_modified: bool ) -> dict[str, Any]: """Create update dict that represents column-values to be updated.""" update_dict = fieldname if isinstance(fieldname, dict) else {fieldname: value} if update_modified: modified = modified or now() modified_by = modified_by or frappe.session.user update_dict.update({"modified": modified, "modified_by": modified_by}) return update_dict def set_single_value( self, doctype: str, fieldname: str | dict, value: str | int | None = None, *, modified=None, modified_by=None, update_modified=True, debug=False, ): """Set field value of Single DocType. :param doctype: DocType of the single object :param fieldname: `fieldname` of the property :param value: `value` of the property Example: # Update the `deny_multiple_sessions` field in System Settings DocType. frappe.db.set_single_value("System Settings", "deny_multiple_sessions", True) """ to_update = self._get_update_dict( fieldname, value, modified=modified, modified_by=modified_by, update_modified=update_modified ) frappe.db.delete( "Singles", filters={"field": ("in", tuple(to_update)), "doctype": doctype}, debug=debug ) singles_data = ((doctype, key, sbool(value)) for key, value in to_update.items()) frappe.qb.into("Singles").columns("doctype", "field", "value").insert(*singles_data).run(debug=debug) frappe.clear_document_cache(doctype, doctype) def get_single_value(self, doctype: str, fieldname: str, cache: bool = True): """Get property of Single DocType. Cache locally by default :param doctype: DocType of the single object whose value is requested :param fieldname: `fieldname` of the property whose value is requested Example: # Get the default value of the company from the Global Defaults doctype. company = frappe.db.get_single_value('Global Defaults', 'default_company') """ if cache and fieldname in self.value_cache[doctype]: return self.value_cache[doctype][fieldname] val = frappe.qb.get_query( table="Singles", filters={"doctype": doctype, "field": fieldname}, fields="value", ).run() val = val[0][0] if val else None df = frappe.get_meta(doctype).get_field(fieldname) if not df: frappe.throw( _("Field {0} does not exist on {1}").format( frappe.bold(fieldname), frappe.bold(doctype), self.InvalidColumnName ) ) val = cast_fieldtype(df.fieldtype, val) self.value_cache[doctype][fieldname] = val return val def get_singles_value(self, *args, **kwargs): """Alias for get_single_value""" return self.get_single_value(*args, **kwargs) def set_value( self, dt: str, dn: FilterValue | dict, field: str, val=None, modified=None, modified_by=None, update_modified=True, debug=False, ): """Set a single value in the database, do not call the ORM triggers but update the modified timestamp (unless specified not to). **Warning:** this function will not call Document events and should be avoided in normal cases. :param dt: DocType name. :param dn: Document name for updating single record or filters for updating many records. :param field: Property / field name or dictionary of values to be updated :param value: Value to be updated. :param modified: Use this as the `modified` timestamp. :param modified_by: Set this user as `modified_by`. :param update_modified: default True. Set as false, if you don't want to update the timestamp. :param debug: Print the query in the developer / js console. """ from frappe.model.utils import is_single_doctype if dn is None or dt == dn: if not is_single_doctype(dt): return from frappe.deprecation_dumpster import deprecation_warning deprecation_warning( "unknown", "v17", "Calling db.set_value on single doctype is deprecated. This behaviour will be removed in future. Use db.set_single_value instead.", ) self.set_single_value( doctype=dt, fieldname=field, value=val, debug=debug, update_modified=update_modified, modified=modified, modified_by=modified_by, ) return to_update = self._get_update_dict( field, val, modified=modified, modified_by=modified_by, update_modified=update_modified ) query = frappe.qb.get_query( table=dt, filters=dn, update=True, validate_filters=True, ) if isinstance(dn, FilterValue): frappe.clear_document_cache(dt, convert_to_value(dn)) else: # No way to guess which documents are modified, clear all of them frappe.clear_document_cache(dt) for column, value in to_update.items(): query = query.set(column, value) query.run(debug=debug) def bulk_update( self, doctype: str, doc_updates: dict, *, chunk_size: int = 100, modified: str | None = None, modified_by: str | None = None, update_modified: bool = True, debug: bool = False, ): """ :param doctype: DocType to update :param doc_updates: Dictionary of key (docname) and values to update :param chunk_size: Number of documents to update in a single transaction :param modified: Use this as the `modified` timestamp. :param modified_by: Set this user as `modified_by`. :param update_modified: default True. Update `modified` and `modified_by` fields :param debug: Print the query in the developer / js console. doc_updates should be in the following format: ```py { "docname1": { "field1": "value1", "field2": "value2", ... }, "docname2": { "field1": "value1", "field2": "value2", ... }, } ``` Note: - Bigger chunk sizes could be less performant. Use appropriate chunk size based on the number of fields to update. """ if not doc_updates: return modified_dict = None if update_modified: modified_dict = self._get_update_dict( {}, None, modified=modified, modified_by=modified_by, update_modified=update_modified ) total_docs = len(doc_updates) iterator = iter(doc_updates.items()) for __ in range(0, total_docs, chunk_size): doc_chunk = dict(itertools.islice(iterator, chunk_size)) self._build_and_run_bulk_update_query(doctype, doc_chunk, modified_dict, debug) @staticmethod def _build_and_run_bulk_update_query( doctype: str, doc_updates: dict, modified_dict: dict | None = None, debug: bool = False ): """ :param doctype: DocType to update :param doc_updates: Dictionary of key (docname) and values to update :param debug: Print the query in the developer / js console. --- doc_updates should be in the following format: ```py { "docname1": { "field1": "value1", "field2": "value2", ... }, "docname2": { "field1": "value1", "field2": "value2", ... }, } ``` --- Query will be built as: ```sql UPDATE `tabItem` SET `status` = CASE WHEN `name` = 'Item-1' THEN 'Close' WHEN `name` = 'Item-2' THEN 'Open' WHEN `name` = 'Item-3' THEN 'Close' WHEN `name` = 'Item-4' THEN 'Cancelled' ELSE `status` end, `description` = CASE WHEN `name` = 'Item-1' THEN 'This is the first task' WHEN `name` = 'Item-2' THEN 'This is the second task' WHEN `name` = 'Item-3' THEN 'This is the third task' WHEN `name` = 'Item-4' THEN 'This is the fourth task' ELSE `description` end WHERE `name` IN ( 'Item-1', 'Item-2', 'Item-3', 'Item-4' ) ``` """ if not doc_updates: return dt = frappe.qb.DocType(doctype) update_query = frappe.qb.update(dt) conditions = {} docnames = list(doc_updates.keys()) for docname, row in doc_updates.items(): for field, value in row.items(): # CASE if field not in conditions: conditions[field] = Case() # WHEN conditions[field].when(dt.name == docname, value) for field in conditions: # ELSE update_query = update_query.set(dt[field], conditions[field].else_(dt[field])) if modified_dict: for column, value in modified_dict.items(): update_query = update_query.set(dt[column], value) update_query.where(dt.name.isin(docnames)).run(debug=debug) def set_global(self, key, val, user="__global"): """Save a global key value. Global values will be automatically set if they match fieldname.""" self.set_default(key, val, user) def get_global(self, key, user="__global"): """Return a global key value.""" return self.get_default(key, user) def get_default(self, key, parent="__default"): """Return default value as a list if multiple or single.""" d = self.get_defaults(key, parent) return (isinstance(d, list) and d[0]) or d @staticmethod def set_default(key, val, parent="__default", parenttype=None): """Sets a global / user default value.""" frappe.defaults.set_default(key, val, parent, parenttype) @staticmethod def add_default(key, val, parent="__default", parenttype=None): """Append a default value for a key, there can be multiple default values for a particular key.""" frappe.defaults.add_default(key, val, parent, parenttype) @staticmethod def get_defaults(key=None, parent="__default"): """Get all defaults""" defaults = frappe.defaults.get_defaults_for(parent) if not key: return defaults if key in defaults: return defaults[key] return defaults.get(frappe.scrub(key)) def begin(self, *, read_only=False): read_only = read_only or frappe.flags.read_only mode = "READ ONLY" if read_only else "" self.sql(f"START TRANSACTION {mode}") def commit(self, *, chain=False): """Commit current transaction. Calls SQL `COMMIT`.""" if self._disable_transaction_control: warnings.warn(message=TRANSACTION_DISABLED_MSG, stacklevel=2) return self.before_rollback.reset() self.after_rollback.reset() self.before_commit.run() if chain: self.sql("commit and chain") else: self.sql("commit") self.begin() self.after_commit.run() def rollback(self, *, save_point=None, chain=False): """`ROLLBACK` current transaction. Optionally rollback to a known save_point.""" if save_point: self.sql(f"rollback to savepoint {save_point}") elif not self._disable_transaction_control: self.before_commit.reset() self.after_commit.reset() self.before_rollback.run() if chain: self.sql("rollback and chain") else: self.sql("rollback") self.begin() self.after_rollback.run() else: warnings.warn(message=TRANSACTION_DISABLED_MSG, stacklevel=2) def savepoint(self, save_point): """Savepoints work as a nested transaction. Changes can be undone to a save point by doing frappe.db.rollback(save_point) Note: rollback watchers can not work with save points. so only changes to database are undone when rolling back to a savepoint. Avoid using savepoints when writing to filesystem.""" self.sql(f"savepoint {save_point}") def release_savepoint(self, save_point): self.sql(f"release savepoint {save_point}") def field_exists(self, dt, fn): """Return true of field exists.""" return self.exists("DocField", {"fieldname": fn, "parent": dt}) def table_exists(self, doctype, cached=True): """Return True if table for given doctype exists.""" return f"tab{doctype}" in self.get_tables(cached=cached) def has_table(self, doctype): return self.table_exists(doctype) def get_tables(self, cached=True): raise NotImplementedError def a_row_exists(self, doctype): """Return True if at least one row exists.""" return frappe.get_all(doctype, limit=1, order_by=None, as_list=True) def exists(self, dt, dn=None, cache=False, *, debug=False): """Return the document name of a matching document, or None. Note: `cache` only works if `dt` and `dn` are of type `str`. ## Examples Pass doctype and docname (only in this case we can cache the result) ``` exists("User", "jane@example.org", cache=True) ``` Pass a dict of filters including the `"doctype"` key: ``` exists({"doctype": "User", "full_name": "Jane Doe"}) ``` Pass the doctype and a dict of filters: ``` exists("User", {"full_name": "Jane Doe"}) ``` """ if dt != "DocType" and dt == dn: # single always exists (!) return dn if isinstance(dt, dict): dt = dt.copy() # don't modify the original dict dt, dn = dt.pop("doctype"), dt return self.get_value(dt, dn, ignore=True, cache=cache, order_by=None, debug=debug) def count(self, dt, filters=None, debug=False, cache=False, distinct: bool = True): """Return `COUNT(*)` for given DocType and filters.""" cache_key = "COUNT(*)" if cache and not filters and cache_key in self.value_cache[dt]: return self.value_cache[dt][cache_key] count = frappe.qb.get_query( table=dt, filters=filters, fields=Count("*"), distinct=distinct, validate_filters=True, ).run(debug=debug)[0][0] if not filters and cache: self.value_cache[dt][cache_key] = count return count def estimate_count(self, doctype: str) -> int: """Get estimated count of total rows in a table.""" raise NotImplementedError @staticmethod def format_date(date): return getdate(date).strftime("%Y-%m-%d") @staticmethod def format_datetime(datetime): if not datetime: return FallBackDateTimeStr return get_datetime(datetime).strftime("%Y-%m-%d %H:%M:%S.%f") def get_creation_count(self, doctype, minutes): """Get count of records created in the last x minutes""" from dateutil.relativedelta import relativedelta from frappe.utils import now_datetime Table = frappe.qb.DocType(doctype) return ( frappe.qb.from_(Table) .select(Count(Table.name)) .where(Table.creation >= now_datetime() - relativedelta(minutes=minutes)) .run()[0][0] ) def get_db_table_columns(self, table) -> list[str]: """Return list of column names from given table.""" key = f"table_columns::{table}" columns = frappe.client_cache.get_value(key) if columns is None: information_schema = frappe.qb.Schema("information_schema") columns = ( frappe.qb.from_(information_schema.columns) .select(information_schema.columns.column_name) .where(information_schema.columns.table_name == table) .run(pluck=True) ) if columns: frappe.cache.set_value(key, columns) return columns def get_table_columns(self, doctype): """Return list of column names from given doctype.""" columns = self.get_db_table_columns("tab" + doctype) if not columns: raise self.TableMissingError("DocType", doctype) return columns def has_column(self, doctype, column): """Return True if column exists in database.""" return column in self.get_table_columns(doctype) def has_index(self, table_name, index_name): raise NotImplementedError def add_index(self, doctype, fields, index_name=None): raise NotImplementedError def add_unique(self, doctype, fields, constraint_name=None): raise NotImplementedError @staticmethod def get_index_name(fields): index_name = "_".join(fields) + "_index" # remove index length if present e.g. (10) from index name return INDEX_PATTERN.sub(r"", index_name) def get_system_setting(self, key): return frappe.get_system_settings(key) def close(self): """Close database connection.""" if self._conn: self._conn.close() self._cursor = None self._conn = None @staticmethod def escape(s, percent=True): """Escape quotes and percent in given string.""" # implemented in specific class raise NotImplementedError from frappe.deprecation_dumpster import is_column_missing as _is_column_missing is_column_missing = staticmethod(_is_column_missing) def get_descendants(self, doctype, name): """Return descendants of the group node in tree""" from frappe.utils.nestedset import get_descendants_of try: return get_descendants_of(doctype, name, ignore_permissions=True) except Exception: # Can only happen if document doesn't exists - kept for backward compatibility return [] def is_missing_table_or_column(self, e): return self.is_missing_column(e) or self.is_table_missing(e) def multisql(self, sql_dict, values=(), **kwargs): current_dialect = self.db_type or "mariadb" query = sql_dict.get(current_dialect) return self.sql(query, values, **kwargs) def delete(self, doctype: str, filters: dict | list | None = None, debug=False, **kwargs): """Delete rows from a table in site which match the passed filters. This does not trigger DocType hooks. Simply runs a DELETE query in the database. Doctype name can be passed directly, it will be pre-pended with `tab`. """ filters = filters or kwargs.get("conditions") query = frappe.qb.get_query( table=doctype, filters=filters, delete=True, validate_filters=True, ) if "debug" not in kwargs: kwargs["debug"] = debug return query.run(**kwargs) def truncate(self, doctype: str): """Truncate a table in the database. This runs a DDL command `TRUNCATE TABLE`. This cannot be rolled back. Doctype name can be passed directly, it will be pre-pended with `tab`. """ return self.sql_ddl(f"truncate `{get_table_name(doctype)}`") def get_last_created(self, doctype): last_record = self.get_all(doctype, ("creation"), limit=1, order_by="creation desc") if last_record: return get_datetime(last_record[0].creation) else: return None def log_touched_tables(self, query, query_type): if query_type not in QUERY_TYPES_FOR_LOG_TOUCHED_TABLES: return # single_word_regex is designed to match following patterns # `tabXxx`, tabXxx and "tabXxx" # multi_word_regex is designed to match following patterns # `tabXxx Xxx` and "tabXxx Xxx" # ([`"]?) Captures " or ` at the beginning of the table name (if provided) # \1 matches the first captured group (quote character) at the end of the table name # multi word table name must have surrounding quotes. # (tab([A-Z]\w+)( [A-Z]\w+)*) Captures table names that start with "tab" # and are continued with multiple words that start with a captital letter # e.g. 'tabXxx' or 'tabXxx Xxx' or 'tabXxx Xxx Xxx' and so on tables = [] for regex in (SINGLE_WORD_PATTERN, MULTI_WORD_PATTERN): tables += [groups[1] for groups in regex.findall(query)] touched_tables = frappe.local.flags.touched_tables if touched_tables is None: touched_tables = set() frappe.local.flags.touched_tables = touched_tables touched_tables.update(tables) def bulk_insert( self, doctype: str, fields: list[str], values: Iterable[Sequence[Any]], ignore_duplicates=False, *, chunk_size=1000, ): """ Insert multiple records at a time :param doctype: Doctype name :param fields: list of fields :params values: iterable of values """ table = frappe.qb.DocType(doctype) query = frappe.qb.into(table).columns(fields) if ignore_duplicates: # Pypika does not have same api for ignoring duplicates if frappe.conf.db_type in ("mariadb", "sqlite"): query = query.ignore() elif frappe.conf.db_type == "postgres": query = query.on_conflict().do_nothing() value_iterator = iter(values) while value_chunk := tuple(itertools.islice(value_iterator, chunk_size)): query.insert(*value_chunk).run() def create_sequence(self, *args, **kwargs): from frappe.database.sequence import create_sequence return create_sequence(*args, **kwargs) def set_next_sequence_val(self, *args, **kwargs): from frappe.database.sequence import set_next_val set_next_val(*args, **kwargs) def get_next_sequence_val(self, *args, **kwargs): from frappe.database.sequence import get_next_val return get_next_val(*args, **kwargs) def get_row_size(self, doctype: str) -> int: """Get estimated max row size of any table in bytes.""" raise NotImplementedError def rename_column(self, doctype: str, old_column_name: str, new_column_name: str): raise NotImplementedError @contextmanager def unbuffered_cursor(self): """Context manager to temporarily use unbuffered cursor. Using this with `as_iterator=True` provides O(1) memory usage while reading large result sets. NOTE: You MUST do entire result set processing in the context, otherwise underlying cursor will be switched and you'll not get complete results. Usage: with frappe.db.unbuffered_cursor(): for row in frappe.db.sql("query with huge result", as_iterator=True): continue # Do some processing. """ raise NotImplementedError @contextmanager def savepoint(catch: type | tuple[type, ...] = Exception): """Wrapper for wrapping blocks of DB operations in a savepoint. as contextmanager: for doc in docs: with savepoint(catch=DuplicateError): doc.insert() as decorator (wraps FULL function call): @savepoint(catch=DuplicateError) def process_doc(doc): doc.insert() """ try: savepoint = "".join(random.sample(string.ascii_lowercase, 10)) frappe.db.savepoint(savepoint) yield # control back to calling function except catch: frappe.db.rollback(save_point=savepoint) else: frappe.db.release_savepoint(savepoint) def get_query_execution_timeout() -> int: """Get execution timeout based on current timeout in different contexts. HTTP requests: HTTP timeout or a default (300) Background jobs: Job timeout Console/Commands: No timeout = 0. Note: Timeout adds 1.5x as "safety factor" """ from rq import get_current_job if not frappe.conf.get("enable_db_statement_timeout"): return 0 # Zero means no timeout, which is the default value in db. timeout = 0 with suppress(Exception): if getattr(frappe.local, "request", None): timeout = frappe.conf.http_timeout or 300 elif job := get_current_job(): timeout = job.timeout return int(cint(timeout) * 1.5) def get_print_sql_flag() -> bool: flag_value = frappe.client_cache.get_value("flag_print_sql") if flag_value is None: flag_value = False frappe.client_cache.set_value("flag_print_sql", flag_value) return flag_value