sqlbind allows to bind parameters in text based raw SQL queries.
>>> q = sqlbind.Dialect.default()
>>> email = '[email protected]'
>>> sql = f'SELECT * FROM users WHERE email = {q/email}'
>>> sql
'SELECT * FROM users WHERE email = ?'
>>> q
['[email protected]']
>>> data = connection.execute(sql, q)Supports all DBAPI parameter styles. Isn't limited by DBAPI compatible drivers and could be used with anything accepting raw SQL query and parameters in some way. For example sqlbind could be used with SQLAlchemy textual queries. Or with clickhouse-driver's non-DBAPI interface.
pip install sqlbind
ORMs are great and could be used effectively for a huge number of tasks. But after many years with SQLAlchemy I've noticed some repeating patterns:
-
It's really not an easy task to decipher complex SQLAlchemy expression back into SQL. Especially when CTEs, sub-queries, nested queries or self-referential queries are involved. It composes quite well but it takes too much effort to write and read SQLAlchemy queries. For novices it could be a hard time to deal with it.
-
Most of reporting queries are big enough already not to be bothered with ORMs and use raw SQL anyway. This kind of SQL often requires dynamic constructs and becomes string fiddling contraption.
-
For a few tasks ORMs bring too much overhead and the only solution is to get down to raw DBAPI connection and raw SQL.
-
(Minor personal grudge, please ignore it) For some ORMs (like Django ORM) your SQL intuition could be useless and requires deep ORM understanding. To the side: sqlalchemy hybrid properties, cough.
It boils down to one thing: from time to time you have to write raw SQL queries. I could highlight 3 types of queries:
-
Fixed queries. They don't contain any parameters. For example
SELECT id, name FROM users ORDER BY registered DESC LIMIT 10. In general fixed queries or fixed query parts compose well and don't require any special treatment. Python's f-strings are enough. -
Static queries. They contain parameters but structure is fully known beforehand. For example
SELECT id, name FROM users WHERE email = :email LIMIT 1. They are also could be composed without large issues, especially for connection drivers supporting named parameters (:param,%(param)s) and accepting dicts as parameters. Although for positional connection drivers (%s,?) composition requires careful parameter tracking and queries could be fragile to change. -
Dynamic queries. Query part presence could depend on parameter value or external condition. For example to provide result on input filter you have to add CTE and corresponding JOIN to a query. Or add filters only for non
Noneinput values. ORMs are effective for composing such queries. Using raw SQL are almost impossible for abstraction and leads to a complex boilerplate heavy code.
Note: here and in following sections I deliberately use simple examples. In real life there is no need to use sqlbind for such kind of queries.
Note: by composing I mean ability to assemble a final query from parts which could be abstracted and reused.
sqlbind tries to address issues with static and dynamic query types. It tracks parameter binds and could help with dynamic query parts.
Some things to consider:
-
sqlbind tries to provide an API for a simple composition of raw SQL. Most operations return string-like objects ready to be inserted in the final query. sqlbind does trivial things and is easy to reason about.
-
There is a large set of functions/methods to address dynamic queries but you haven't use it inline in a single query string. You could use variables to keep query parts and stitch resulted SQL from these parts.
-
This README misses large portions of API. Feel free to explore doc strings with examples of fully strictly type-hinted sqlbind's source code!
General use case looks like:
# a global alias to a dialect used by connection backend, see `sqlbind.Dialect`
QParams = sqlbind.Dialect.some_dialect
def get_my_data(value1, value2):
# Construct empty fresh sqlbind.QueryParams
q = QParams()
# Use `q` to bind parameter values in SQL string.
sql = f'SELECT * FROM table WHERE field1 = {q/value1} AND field2 > {q/value2}'
# Pass query and parameters into connection's execute.
return get_connection().execute(sql, q).fetchall()For queries or query parts with a known structure the most simple way to bind a parameter is to
use bind operator /:
>>> date = "2023-01-01"
>>> q = sqlbind.Dialect.default()
>>> f'SELECT * FROM users WHERE registered > {q/date}'
'SELECT * FROM users WHERE registered > ?'
>>> q
['2023-01-01']Or for named style parameters:
>>> date = "2023-01-01"
>>> q = sqlbind.Dialect.default_named()
>>> f'SELECT * FROM users WHERE registered > {q/date}'
'SELECT * FROM users WHERE registered > :p0'
>>> q
{'p0': '2023-01-01'}There is no any magic. Bind operator returns a string with a placeholder for a
corresponding dialect and adds parameter's value to q object. That's all.
q object is inherited from a dict or a list depending from a used
dialect.
>>> value = 10
>>> q = sqlbind.Dialect.default()
>>> q/value
'?'
>>> q
[10]Note: there is no much value in sqlbind if you have only static queries and use connection backends accepting named parameters.
Here begins a fun part. We can't use simple binds for dynamic queries. For example we have a function returning recently registered users:
def get_fresh_users(registered_since: datetime):
q = QParams() # an alias to some dialect to construct sqlbind.QueryParams instance
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since}
ORDER BY registered
'''
return connection.execute(sql, q)And later there is a new requirement for the function. It should return only enabled or only disabled users if corresponding argument is passed.
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
if enabled is not None:
enabled_filter = f' AND enabled = {q/enabled}'
else:
enabled_filter = ''
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since} {enabled_filter}
ORDER BY registered
'''
return connection.execute(sql, q)It looks almost pretty. See how q/enabled helped to track additional parameter.
But you can predict where we are going. Another one or two
additional filters and it would be a complete mess. Take note how WHERE lost AND
between two filters.
In reality bind operator / is a sugar on top of generic sqlbind's API to
bind parameters via q-templates.
>>> q = sqlbind.Dialect.default()
>>> q('field BETWEEN {} AND {}', 10, 20)
'field BETWEEN ? AND ?'
>>> q
[10, 20]QueryParams q object is also a callable accepting a template with {}
placeholders and following parameters to substitute. q/value is same as calling
q('{}', value)
>>> q/10
'?'
>>> q('{}', 10)
'?'You could use q-templates to bind parameters in complex SQL expressions.
q.cond could render a q-template as an empty string based on some condition.
>>> enabled = True
>>> q.cond(enabled is not None, ' AND enabled = {}', enabled)
' AND enabled = ?'
>>> enabled = None
>>> q.cond(enabled is not None, ' AND enabled = {}', enabled)
''q.cond is a generic form. To remove a repetition (enabled is not None/enabled) when value is used both in a condition and as a parameter
value there are two helpers for most common cases:
q.not_none: to check value is not None.q.truthy: to check value's trueness (bool(value) is True).not_emptycould be used as an alias totruthy.
>>> enabled = True
>>> q.not_none(' AND enabled = {}', enabled)
' AND enabled = ?'
>>> enabled = None
>>> q.not_none(' AND enabled = {}', enabled)
''Let's try it in the function:
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
enabled_filter = q.not_none(' AND enabled = {}', enabled)
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since} {enabled_filter}
ORDER BY registered
'''
return connection.execute(sql, q)Hmm. But really nothing was changed. You could write previous code with ternary if/else and it would look the same from semantic standpoint. May be use it inline?
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since}
{q.not_none(' AND enabled = {}', enabled)}
ORDER BY registered
'''
return connection.execute(sql, q)Ugh. Abomination, to say at least.
ANDin the middle of a cryptic expression.q.not_noneandenabledare far away and it's not obvious they are connected- expression is too long and noisy
Let's tackle issues bit by bit.
Prependers could render non-empty inputs with corresponding prefixes and empty string otherwise.
>>> AND_('field1 > 1', 'field2 < 1')
'AND field1 > 1 AND field2 < 1'
>>> OR_('field1 > 1', 'field2 < 1')
'OR field1 > 1 OR field2 < 1'
>>> AND_(q.not_none('enabled = {}', True))
'AND enabled = ?'
>>> AND_(q.not_none('enabled = {}', None))
''Our function with prependers:
from sqlbind import AND_
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since}
{AND_(q.not_none('enabled = {}', enabled))}
ORDER BY registered
'''
return connection.execute(sql, q)At least AND is almost on it's place in SQL structure.
Conditional markers sqlbind.not_none/sqlbind.truthy/sqlbind.cond allows to tie conditionals
with a value via / operator:
>>> q('enabled = {}', sqlbind.not_none/10)
'enabled = ?'
>>> q('enabled = {}', sqlbind.not_none/None)
''Conditional markers return value itself or special UNDEFINED object. UNDEFINED parameters force expressions to be rendered as empty strings.
sqlbind.not_none returns UNDEFINED if value is None:
>>> sqlbind.not_none/10
10
>>> sqlbind.not_none/None is sqlbind.UNDEFINED
Truesqlbind.truthy or sqlbind.not_empty returns UNDEFINED if bool(value) != True:
>>> sqlbind.truthy/10
10
>>> sqlbind.not_empty/10
10
>>> sqlbind.truthy/0 is sqlbind.UNDEFINED
Truesqlbind.cond returns UNDEFINED if condition is False:
>>> sqlbind.cond(True)/10
10
>>> sqlbind.cond(False)/10 is sqlbind.UNDEFINED
TrueNote: sqlbind.cond is almost always awkward to use inline in real life and exists largely for symmetry with q.cond.
Rewritten function:
from sqlbind import AND_, not_none
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since}
{AND_(q('enabled = {}', not_none/enabled))}
ORDER BY registered
'''
return connection.execute(sql, q)Almost there. May be there is a way to reduce number of quotes inside AND_?
q-expressions allow to generate templated results with infix operators.
Any unknown attribute access to q object returns QExpr which has str
conversion as an attribute name:
>>> str(q.field)
'field'
>>> str(q.table.field)
'table.field'q has a number of attributes itself those names could conflict with existing
DB tables/columns. To resolve conflicts you could use q._. (stare) expression:
>>> str(q._.cond)
'cond'Real DB tables/columns could use quite peculiar names. You could youse q._
(pirate) expression to construct QExpr from any string:
>>> str(q._('"weird table"."weird column"'))
'"weird table"."weird column"'QExpr object knows about parent q object and defines a set of infix operators
allowing to bind a right value:
>>> q.field > 10
'field > ?'
>>> q.table.field == 20
'table.field = ?'
>>> q._.table.field == None
'table.field IS NULL'
>>> q._('"my column"') != None
'"my column" IS NOT NULL'
>>> q.field <= not_none/None # conditional marks also work!
''
>>> q.field.IN(not_none/[10]) # BTW sqlbind has workaround for SQLite to deal with arrays in IN
'field IN ?'It could look like a hack and feel ORM-ish but there is no any expression trees and tree compilation passes. q-expressions are immediately rendered as strings and simple to reason about.
Also set of operations is really small it includes only comparisons and QExpr.IN.
Let's use q-expressions with the function:
from sqlbind import AND_, not_none
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
sql = f'''\
SELECT * FROM users
WHERE registered > {q/registered_since}
{AND_(q.enabled == not_none/enabled)}
ORDER BY registered
'''
return connection.execute(sql, q)I have no any other tricks. It's the final inline version. I can't make it
more pretty or readable. It's true, inline expressions looks a bit noisy and to
make it manageable try to extract as much logic and use only not_none conditional marker.
IMHO instead of
>>> now = None
>>> show_only_enabled = True
>>> f'SELECT * FROM users WHERE registered > {q/((now or datetime.utcnow()) - timedelta(days=30))} {AND_(q.enabled == cond(show_only_enabled)/1)}'
'SELECT * FROM users WHERE registered > ? AND enabled = ?'please consider to use:
>>> now = None
>>> show_only_enabled = True
>>> registered_since = (now or datetime.utcnow()) - timedelta(days=30)
>>> enabled = 1 if show_only_enabled else None
>>> f'SELECT * FROM users WHERE registered > {q/registered_since} {AND_(q.enabled == not_none/enabled)}'
'SELECT * FROM users WHERE registered > ? AND enabled = ?'Also there is a possibility to construct filters out of line via WHERE
prepender.
It could be useful to extract filters outside of f-strings and use sqlbind.WHERE
prepender. It can help with readability of long complex filters.
from sqlbind import not_none, WHERE
def get_fresh_users(registered_since: datetime, enabled: Optional[bool] = None):
q = QParams()
filters = [
q.registered > registered_since,
q.enabled == not_none/enabled,
]
sql = f'SELECT * FROM users {WHERE(*filters)} ORDER BY registered'
return connection.execute(sql, q)There are also other prependers: WITH, LIMIT, OFFSET, GROUP_BY,
ORDER_BY, SET. They all omit empty parts or are rendered as
empty string if all parts are empty.
Also you could use & operator to join filters to assemble condition expression without a list:
>>> filters = (q.registered > '2023-01-01') & (q.enabled == not_none/True)
>>> WHERE(filters)
'WHERE (registered > ? AND enabled = ?)'β "Wait a minute. How does it work? You said there is no expression trees and compilation! And all operations return strings!"
Well, technically they are strings. Almost all methods and functions return sqlbind.Expr. It's a very shallow
descendant of str with only __or__, __and__ and __invert__ overrides.
>>> q('enabled') & q('registered')
'(enabled AND registered)'
>>> type(q('enabled'))
<class 'sqlbind.Expr'>
>>> type(q.enabled == True)
<class 'sqlbind.Expr'>All Expr instances could be composed with &, | and ~ (negate) operations.
Sadly due to python's' precedence rules you have to wrap expressions into
additional parens to make it work.
It's a matter of preference and team code agreements. Personally I don't see anything criminal in inline expressions. But it could be a huge red flag for other person and it's ok. sqlbind gives a choice to use inline or out of line approach.
But take a note. For positional dialects (like qmark style) out of line rendering has a major drawback. You should take care on part ordering. Binding and part usage should be synchronised. For example:
>>> q = sqlbind.Dialect.default()
>>> filter1 = q.registered > '2023-01-01'
>>> filter2 = q.enabled == 1
>>> f'SELECT * FROM users WHERE {filter2} AND {filter1}'
'SELECT * FROM users WHERE enabled = ? AND registered > ?'
>>> q # parameter ordering mismatches placeholders
['2023-01-01', 1]It's a largely artificial example but for complex queries composed from multiple parts it could be an issue. To reduce chance you could abstract composition parts in a way to contain bindings and SQL construction in one go to be fully synchronised.
BTW, you could already noticed but out of line variants of get_fresh_users
from Dynamic queries and Conditionals have
the same ordering bug: inline and out of line approaches mix quite bad. Always
use named style Dialect if your connection backend allows it.