Named rewrites are used to keep track of the user intent for operations and macros that are defined in terms of other basic operations.
Named rewrites are used to keep track of the user intent for operations and
macros that are defined in terms of other basic operations. For example, :avg
is not available as a basic aggregate type, it is a rewrite to
query,:sum,query,:count,:div
. However, for the user it is better if we can
show query,:avg
when dumping the expression as a string.
Name of the operation, e.g., avg
.
Expression that is displayed to the user when creating the expression string.
Expression that is evaluated.
Evaluation context for the initial creation time. This context is used to
re-evaluate the rewrite using the original context if the overall expression
is rewritten (Expr.rewrite()
) later.
Compute estimated percentile values using counts for well known buckets.
Compute estimated percentile values using counts for well known buckets. See spectator
PercentileBuckets for more information. The input will be grouped by the percentile
key
with each key value being the bucket index. The output will be one line per requested
percentile.
Input data expression. The value should be a sum or group by. The group by list should
include the 'percentile' key. If using the :percentiles
word to construct the instance
then other aggregate types, such as max, will automatically be converted to sum and the
percentile
key will be added into the group by clause.
List of percentiles to compute. Each value should be in the range [0.0, 100.0].
Same as [Random], but allows the user to specify a seed to vary the input.
Same as [Random], but allows the user to specify a seed to vary the input. This allows multiple sample lines to be produced with different values.
Generate a time series that appears to be random noise for the purposes of experimentation and generating sample data.
Generate a time series that appears to be random noise for the purposes of experimentation and generating sample data. To ensure that the line is deterministic and reproducible it actually is based on a hash of the timestamp.
(Since version ) see corresponding Javadoc for more information.