docs: clarify read_consistency_interval scope and cost

This commit is contained in:
Brendan Clement
2026-05-22 20:40:33 -07:00
parent cbdad17f34
commit e363e1661f
3 changed files with 25 additions and 13 deletions

View File

@@ -24,15 +24,19 @@ mod util;
#[napi(object)]
#[derive(Debug)]
pub struct ConnectionOptions {
/// (For LanceDB OSS only): The interval, in seconds, at which to check for
/// updates to the table from other processes. If None, then consistency is not
/// checked. For performance reasons, this is the default. For strong
/// consistency, set this to zero seconds. Then every read will check for
/// updates from other processes. As a compromise, you can set this to a
/// non-zero value for eventual consistency. If more than that interval
/// has passed since the last check, then the table will be checked for updates.
/// Note: this consistency only applies to read operations. Write operations are
/// The interval, in seconds, at which to check for updates to the table
/// from other processes. If None, then consistency is not checked. For
/// performance reasons, this is the default. For strong consistency, set
/// this to zero seconds. Then every read will check for updates from other
/// processes. As a compromise, you can set this to a non-zero value for
/// eventual consistency. If more than that interval has passed since the
/// last check, then the table will be checked for updates. Note: this
/// consistency only applies to read operations. Write operations are
/// always consistent.
///
/// Stronger consistency is not free. The smaller the interval, the more
/// often each read pays the cost of checking for updates against object
/// storage, raising per-read latency and cost.
pub read_consistency_interval: Option<f64>,
/// (For LanceDB OSS only): configuration for object storage.
///

View File

@@ -94,7 +94,6 @@ def connect(
host_override: str, optional
The override url for LanceDB Cloud.
read_consistency_interval: timedelta, default None
(For LanceDB OSS only)
The interval at which to check for updates to the table from other
processes. If None, then consistency is not checked. For performance
reasons, this is the default. For strong consistency, set this to
@@ -104,6 +103,10 @@ def connect(
the last check, then the table will be checked for updates. Note: this
consistency only applies to read operations. Write operations are
always consistent.
Stronger consistency is not free. The smaller the interval, the more
often each read pays the cost of checking for updates against object
storage, raising per-read latency and cost.
client_config: ClientConfig or dict, optional
Configuration options for the LanceDB Cloud HTTP client. If a dict, then
the keys are the attributes of the ClientConfig class. If None, then the
@@ -344,7 +347,6 @@ async def connect_async(
host_override: str, optional
The override url for LanceDB Cloud.
read_consistency_interval: timedelta, default None
(For LanceDB OSS only)
The interval at which to check for updates to the table from other
processes. If None, then consistency is not checked. For performance
reasons, this is the default. For strong consistency, set this to
@@ -354,6 +356,10 @@ async def connect_async(
the last check, then the table will be checked for updates. Note: this
consistency only applies to read operations. Write operations are
always consistent.
Stronger consistency is not free. The smaller the interval, the more
often each read pays the cost of checking for updates against object
storage, raising per-read latency and cost.
client_config: ClientConfig or dict, optional
Configuration options for the LanceDB Cloud HTTP client. If a dict, then
the keys are the attributes of the ClientConfig class. If None, then the

View File

@@ -824,9 +824,11 @@ impl ConnectBuilder {
/// This only affects read operations. Write operations are always
/// consistent.
///
/// For LanceDB Cloud and Enterprise, the interval is sent on every read as
/// an `x-lancedb-min-timestamp` freshness header so the server's cache
/// honors the same semantics.
/// # Cost
///
/// Stronger consistency is not free. The smaller the interval, the more
/// often each read pays the cost of checking for updates against object
/// storage, raising per-read latency and cost.
pub fn read_consistency_interval(
mut self,
read_consistency_interval: std::time::Duration,