Skip to content

Optimize state operations #891

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
May 23, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 11 additions & 6 deletions quixstreams/state/base/transaction.py
Original file line number Diff line number Diff line change
Expand Up @@ -388,7 +388,6 @@ def _get_bytes(

return cached

@validate_transaction_status(PartitionTransactionStatus.STARTED)
def set(self, key: K, value: V, prefix: bytes, cf_name: str = "default") -> None:
"""
Set value for the key.
Expand All @@ -404,9 +403,8 @@ def set(self, key: K, value: V, prefix: bytes, cf_name: str = "default") -> None
self._status = PartitionTransactionStatus.FAILED
raise

self.set_bytes(key, value_serialized, prefix, cf_name=cf_name)
self._set_bytes(key, value_serialized, prefix, cf_name=cf_name)

@validate_transaction_status(PartitionTransactionStatus.STARTED)
def set_bytes(
self, key: K, value: bytes, prefix: bytes, cf_name: str = "default"
) -> None:
Expand All @@ -417,10 +415,17 @@ def set_bytes(
:param value: value
:param cf_name: column family name
"""
try:
if not isinstance(value, bytes):
raise StateSerializationError("Value must be bytes")
if not isinstance(value, bytes):
self._status = PartitionTransactionStatus.FAILED
raise StateSerializationError("Value must be bytes")

self._set_bytes(key=key, value=value, prefix=prefix, cf_name=cf_name)

@validate_transaction_status(PartitionTransactionStatus.STARTED)
def _set_bytes(
self, key: K, value: bytes, prefix: bytes, cf_name: str = "default"
) -> None:
try:
key_serialized = self._serialize_key(key, prefix=prefix)
self._update_cache.set(
key=key_serialized,
Expand Down
25 changes: 9 additions & 16 deletions quixstreams/state/rocksdb/windowed/transaction.py
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@
from .partition import WindowedRocksDBStorePartition


class WindowedRocksDBPartitionTransaction(PartitionTransaction[bytes, dict]):
class WindowedRocksDBPartitionTransaction(PartitionTransaction[bytes, Any]):
def __init__(
self,
partition: "WindowedRocksDBStorePartition",
Expand Down Expand Up @@ -108,13 +108,11 @@ def keys(self, cf_name: str = "default") -> Iterable[Any]:
yield key

def get_latest_timestamp(self, prefix: bytes) -> int:
return self._get_timestamp(
prefix=prefix, cache=self._latest_timestamps, default=0
)
return self._get_timestamp(prefix=prefix, cache=self._latest_timestamps) or 0

def get_latest_expired(self, prefix: bytes) -> int:
return self._get_timestamp(
prefix=prefix, cache=self._last_expired_timestamps, default=0
return (
self._get_timestamp(prefix=prefix, cache=self._last_expired_timestamps) or 0
)

def get_window(
Expand Down Expand Up @@ -315,9 +313,7 @@ def expire_all_windows(
:param delete: If True, expired windows will be deleted.
:param collect: If True, values will be collected into windows.
"""
last_expired = self._get_timestamp(
prefix=b"", cache=self._last_expired_timestamps, default=0
)
last_expired = self.get_latest_expired(prefix=b"")

to_delete: set[tuple[bytes, int, int]] = set()
collected = []
Expand Down Expand Up @@ -526,18 +522,15 @@ def _get_items(
# Sort and deserialize items merged from the cache and store
return sorted(merged_items.items(), key=lambda kv: kv[0], reverse=backwards)

def _get_timestamp(
self, cache: TimestampsCache, prefix: bytes, default: Any = None
) -> int:
cached_ts = cache.timestamps.get(prefix)
if cached_ts is not None:
return cached_ts
def _get_timestamp(self, cache: TimestampsCache, prefix: bytes) -> Optional[int]:
if prefix in cache.timestamps:
# Return the cached value if it has been set at least once
return cache.timestamps[prefix]

stored_ts = self.get(
key=cache.key,
prefix=prefix,
cf_name=cache.cf_name,
default=default,
)
if stored_ts is not None and not isinstance(stored_ts, int):
raise ValueError(f"invalid timestamp {stored_ts}")
Expand Down