Dealing with Predicate Expression Filters in Aerospike REST Client (Part 2)

In the previous part we spoke about using simple Predicate Expression Filters with the Aerospike REST Client. In this post I want to dive deeper and show how special filter operators could come in handy.

LAST_UPDATE
Asserts the last update time before processing the transaction.

Parameters:

• comparison operator
• time in seconds since the epoch time (1970-01-01)

Example:
LAST_UPDATE(>=, 1577880000)

Java equivalent:

PredExp.recLastUpdate()
PredExp.integerValue(1577880000000000000) //in nano
PredExp.integerGreaterEq()

VOID_TIME
Asserts the record’s expiration time before processing the transaction.

Parameters:

• comparison operator
• time in seconds since the epoch time (1970-01-01)

Example:
VOID_TIME(>=, 1577880000)

Java equivalent:

PredExp.recVoidTime()
PredExp.integerValue(1577880000000000000) //in nano
PredExp.integerGreaterEq()

DIGEST_MODULO
Asserts a modulo against the record’s digest. The 160 bits (20 Bytes) digest field is generated by hashing the set name and the user key using the RipeMD-160 hashing function.

Parameters:

• modulo divisor
• comparison operator
• expected remainder value

Example for digest(key) % 3 == 1:
DIGEST_MODULO(3, ==, 1)

Java equivalent:

PredExp.recDigestModulo(3)
PredExp.integerValue(1)
PredExp.integerEqual()

STRING_REGEX
Asserts that the regular expression matches the string valued bin.

Parameters:

• string bin name
• regular expression

Example:
STRING_REGEX(str, [0-9]*)

Java equivalent:

PredExp.stringBin("str")
PredExp.stringValue("[0-9]*")
PredExp.stringRegex(RegexFlag.NONE)

LIST_ITERATE_OR
Creates a predicate expression to match any item in a list value bin.

Parameters:

• list bin name
• comparison operator
• value

Example:
Find records where any list item v = "hello" in list bin x.
LIST_ITERATE_OR(x, ==, hello)

Java equivalent:

PredExp.stringVar("v")
PredExp.stringValue("hello")
PredExp.stringEqual()
PredExp.listBin("x")
PredExp.listIterateOr("v")

LIST_ITERATE_AND
Creates a predicate expression to match all items in a list value bin.

Parameters:

• list bin name
• comparison operator
• value

Example:
Find records where all list elements v != "world" in list bin x.
LIST_ITERATE_AND(x, !=, world)

Java equivalent:

PredExp.stringVar("v")
PredExp.stringValue("world")
PredExp.stringUnequal()
PredExp.listBin("x")
PredExp.listIterateAnd("v")

MAPKEY_ITERATE_OR
Creates a predicate expression to match any map key.

Parameters:

• map bin name
• comparison operator
• value

Example:
Find records where any map key k = 7 in map bin m.
MAPKEY_ITERATE_OR(m, ==, 7)

Java equivalent:

PredExp.integerVar("k")
PredExp.integerValue(7)
PredExp.integerEqual()
PredExp.mapBin("m")
PredExp.mapKeyIterateOr("k")

MAPVAL_ITERATE_OR
Creates a predicate expression to match any map value.

Parameters:

• map bin name
• comparison operator
• value

Example:
Find records where any map value v > 100 in map bin m.
MAPVAL_ITERATE_OR(m, >, 100)

Java equivalent:

PredExp.integerVar("v")
PredExp.integerValue(100)
PredExp.integerGreater()
PredExp.mapBin("m")
PredExp.mapValIterateOr("v")

MAPKEY_ITERATE_AND
Creates a predicate expression to match all map keys.

Parameters:

• map bin name
• comparison operator
• value

Example:
Find records where all map keys k < 5 in map bin m.
MAPKEY_ITERATE_AND(m, <, 5)

Java equivalent:

PredExp.integerVar("k")
PredExp.integerValue(5)
PredExp.integerLess()
PredExp.mapBin("m")
PredExp.mapKeyIterateAnd("k")

MAPVAL_ITERATE_AND
Creates a predicate expression to match all map values.

Parameters:

• map bin name
• comparison operator
• value

Example:
Find records where all map values v > 500 in map bin m.
MAPVAL_ITERATE_AND(m, >, 500)

Java equivalent:

PredExp.integerVar("v")
PredExp.integerValue(500)
PredExp.integerGreater()
PredExp.mapBin("m")
PredExp.mapValIterateAnd("v")

At the end you can glue those with logical operators (and | or) mixing with simple filters, Base64-encode the whole expression and send it as a GET parameter.

KVS_URL = 'http://localhost:8080/v1/kvs'

# In this example I put all metadata checks first, because Aerospike doesn't have an
# optimizer that knows how to shuffle predicates around more efficiently. This way, if the
# data is on an SSD, we can save on reading the record.
exp = b'(c >= 11 and not c < 20) or LAST_UPDATE(>=, 1577880000) or DIGEST_MODULO(3, ==, 1)'
encoded_exp = base64.b64encode(exp)

predexp_uri = '{base}/{ns}/{setname}/{userkey}'.format(
    base=KVS_URL, ns=namespace, setname=setname, userkey=userkey)

response = requests.get(predexp_uri, {"predexp": encoded_exp})
print(response.json())

To summarize, we’ve seen that Predicate Expression Filters are a very powerful tool and learned how to use them with the Aerospike REST Client.

If you still haven’t tried Aerospike REST Client and want to do so, you can find it here.