select DivAirportIDs,
arrayIndexOfInt(DivAirportIDs, 14683) AS index
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
limit 5
ADD
This section contains reference documentation for the ADD function.
Sum of at least two values
Signature
ADD(col1, col2, col3...)
Usage Examples
These examples are based on the .
homeRuns
baseOnBalls
total
ago
This section contains reference documentation for the ago function.
Return time as epoch millis before the given period (in ISO-8601 duration format).
Examples:
"PT20.345S" -- parses as "20.345 seconds"
"PT15M" -- parses as "15 minutes" (where a minute is 60 seconds)
"PT10H" -- parses as "10 hours" (where an hour is 3600 seconds)
"P2D" -- parses as "2 days" (where a day is 24 hours or 86400 seconds)
"P2DT3H4M" -- parses as "2 days, 3 hours and 4 minutes"
"P-6H3M" -- parses as "-6 hours and +3 minutes"
"-P6H3M" -- parses as "-6 hours and -3 minutes"
"-P-6H+3M" -- parses as "+6 hours and -3 minutes"
Signature
ago()
Usage Examples
oneDayAgo
This function is typically used in the predicate to filter on timestamps for recent data. e.g. filter data on recent 1 day.
arrayConcatInt
This section contains reference documentation for the arrayConcatInt function.
Concatenates two arrays of ints.
Signature
arrayConcatInt('colName1', 'colName2')
arrayDistinctInt
This section contains reference documentation for the arrayDistinctInt function.
Returns unique values in an array of ints.
Signature
arrayDistinctInt('colName')
CHR
This section contains reference documentation for the CHR function.
the character corresponding to the Unicode codepoint
Signature
CHR(codepoint)
ABS
This section contains reference documentation for the abs function.
Absolute of a value
Signature
ABS(col1)
DISTINCTAVGMV
This section contains reference documentation for the DISTINCTAVGMV function.
Returns the average of distinct row values in a group
Signature
DISTINCTAVGMV(colName)
ceil
This section contains reference documentation for the CEIL function.
Rounded up to the nearest integer.
Signature
CEIL(col1)
count
This section contains reference documentation for the count function.
Get the count of rows in a group
Signature
COUNT(colName)
select homeRuns, baseOnBalls, ADD(homeRuns, baseOnBalls) AS total
from baseballStats
WHERE teamID = 'ML1'
AND yearID = 1956
AND playerName = 'Henry Louis'
SELECT *
FROM tableName
WHERE tsInMillis > ago('P1D')
select DivWheelsOffs,
arrayConcatInt(DivWheelsOffs, DivWheelsOns) AS concatIds
from airlineStats
WHERE arraylength(DivWheelsOffs) >= 2
limit 5
select DivAirportIDs,
arrayDistinctInt(DivAirportIDs) AS unique
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
limit 5
arrayContainsInt
This section contains reference documentation for the arrayContainsInt function.
Checks if int value exists in array.
Signature
arrayContainsInt('colName', valueToFind)
Usage Examples
These examples are based on the .
DivAirportIDs
containsValue
AVGMV
This section contains reference documentation for the AVGMV function.
Get the avg of values in a group
Signature
AVGMV(colName)
Usage Examples
These examples are based on the .
value
EXPR_MIN / EXPR_MAX
This section contains reference documentation for the EXPR_MIN and EXPR_MAX function.
This function scans the given dataset to identify the maximum and minimum values in the specified measuring columns. Once these extreme values (the maxima and minima) are found, the function locates the corresponding entries in the projection column. These entries are associated with the rows where the extreme values were found in the measuring columns. The function then returns these projection column values, providing a way to link the extreme measurements with their corresponding data in another part of the dataset.
Find the user with maximum activity. If there are multiple users, break the tie with their last_activity_date. If still a tie, break with user_id. And project user_id.
More useful is that this multiple such aggregation function can be used with GROUP BY
Note:
In cases where multiple rows share the same extreme values in the measuring columns, all such rows will be returned by the function.
If the goal is to project multiple different columns that correspond to the same set of measuring columns, you can achieve this by invoking the function multiple times, each time specifying a different projection column.
This impl does not work with AS clause (e.g. SELECT exprmin(longCol, doubleCol) AS exprmin won't work)
For more detailed examples, see:
arrayDistinctString
This section contains reference documentation for the arrayDistinctString function.
Returns unique values in an array of strings.
Signature
arrayDistinctString('colName')
Usage Examples
These examples are based on the .
DivTailNums
unique
arraySortInt
This section contains reference documentation for the arraySortInt function.
Sorts array of ints.
Signature
arraySortInt('colName')
Usage Examples
These examples are based on the .
DivAirportIDs
sortedIds
arrayConcatString
This section contains reference documentation for the arrayConcatString function.
Concatenates two arrays of strings.
Signature
arrayConcatString('colName1', 'colName2')
Usage Examples
These examples are based on the .
DivTailNums
concatIds
arrayUnionInt
This section contains reference documentation for the arrayUnionInt function.
Create a union of two arrays of ints.
Signature
arrayUnionInt('colName1', 'colName2')
Usage Examples
These examples are based on the .
DivWheelsOffs
DivWheelsOns
unionIds
arrayReverseString
This section contains reference documentation for the arrayReverseString function.
Reverses array of strings.
Signature
arrayReverseString('colName')
Usage Examples
These examples are based on the .
FlightNum
reversedAirports
RandomAirports
DISTINCTSUMMV
This section contains reference documentation for the DISTINCTSUMMV function.
Returns the sum of the distinct row values in a group
Signature
DISTINCTSUMMV(colName)
Usage Examples
These examples are based on the .
VALUE
arrayRemoveString
This section contains reference documentation for the arrayRemoveString function.
Removes value from array of strings.
Signature
arrayRemoveString('colName', value)
Usage Examples
These examples are based on the .
DivAirportIDs
value
DIV
This section contains reference documentation for the DIV function.
Quotient of two values
Signature
DIV(col1, col2)
Usage Examples
These examples are based on the .
homeRuns
numberOfGames
total
codepoint
This section contains reference documentation for the CODEPOINT function.
the Unicode codepoint of the first character of the string
Signature
CODEPOINT(col)
Usage Examples
value
length
This section contains reference documentation for the length function.
calculate length of the string
Signature
LENGTH(col)
Usage Examples
value
arrayConcatLong
This section contains reference documentation for the arrayConcatLong function.
Concatenates two arrays of longs.
Signature
arrayConcatLong('colName1', 'colName2')
Usage Examples
This example assumes the multiValueTable columns mvCol1 and mvCol2 are both of type LONG with singleValueField in the table schema set to false.
ln
This section contains reference documentation for the ln function.
Natural log of value i.e. ln(col1)
Signature
LN(col1)
Usage Examples
value
value
minmaxrange
This section contains reference documentation for the minmaxrange function.
Returns the max - min value in a group
Signature
MINMAXRANGE(colName)
Usage Examples
These examples are based on the .
value
now
This section contains reference documentation for the now function.
Return current time as epoch millis.
Signature
now()
Usage Examples
now
This function is typically used in predicate to filter on timestamp for recent data. e.g. filter data on recent 1 day(86400 seconds)
lower
This section contains reference documentation for the lower function.
Converts string to lower case.
Signature
LOWER(col)
Usage Examples
name
DISTINCT
This section contains reference documentation for the DISTINCT function.
Returns the distinct row values in a group
Signature
DISTINCT(colName)
concat
This section contains reference documentation for the concat function.
Concatenate two input strings using the seperator
Signature
CONCAT(col1, col2, seperator)
DISTINCTAVG
This section contains reference documentation for the DISTINCTAVG function.
Returns the average of distinct row values in a group
Signature
DISTINCTAVG(colName) or avg(distinct col)
arrayContainsString
This section contains reference documentation for the arrayContainsString function.
Checks if string value exists in array.
Signature
arrayContainsString('colName', valueToFind)
arraySliceInt
This section contains reference documentation for the arraySliceInt function.
Returns the values in the array between the start and end positions.
Signature
arraySliceInt('colName', start, end)
day
This section contains reference documentation for the day function.
Returns the day of the month from the given epoch millis in UTC or specified timezone. The value ranges from 1 to 31.
Signature
day(tsInMillis)
day(tsInMillis, timeZoneId)
arrayIndexOfString
This section contains reference documentation for the arrayIndexOfString function.
Finds the last index of the given value in the array starting at the given index.
Signature
arrayIndexOfString('colName', valueToFind)
DISTINCTCOUNTBITMAP
This section contains reference documentation for the DISTINCTCOUNTBITMAP function.
Returns the count of distinct row values in a group. This function is accurate for INT column, but approximate for other cases where hash codes are used in distinct counting and there may be hash collisions.
For accurate distinct counting on all column types, see .
Signature
DISTINCTCOUNTBITMAP(colName)
DISTINCTCOUNTHLLMV
This section contains reference documentation for the DISTINCTCOUNTBITMAPMV function.
Returns the count of distinct row values in a group. This function is accurate for an INT or dictionary encoded column, but approximate for other cases where hash codes are used in distinct counting and there may be hash collision.
Signature
DISTINCTCOUNTBITMAPMV(colName)
DISTINCTCOUNTRAWHLL
This section contains reference documentation for the DISTINCTCOUNTRAWHLL function.
Returns HLL response serialized as string. The serialized HLL can be converted back into an HLL and then aggregated with other HLLs. A common use case may be to merge HLL responses from different Pinot tables, or to allow aggregation after client-side batching.
Signature
DISTINCTCOUNTRAWHLL(colName, log2m)
dayOfYear
This section contains reference documentation for the dayOfYear function.
Returns the day of the year from the given epoch millis in UTC or specified timezone. The value ranges from 1 to 366.
Signature
dayOfYear(tsInMillis)
dayOfYear(tsInMillis, timeZoneId)
arrayRemoveInt
This section contains reference documentation for the arrayRemoveInt function.
Removes value from array of ints.
Signature
arrayRemoveInt('colName', value)
COVAR_SAMP
This section contains reference documentation for the COVAR_SAMP function.
Returns the sample covariance between of 2 numerical columns.
Signatures
COVAR_SAMP(col1, col2) -> double
COUNTMV
This section contains reference documentation for the COUNTMV function.
Get the count of rows in a group
Signature
COUNTMV(colName)
exp
This section contains reference documentation for the exp function.
Euler’s number(e) raised to the power of col.
Signature
EXP(col1)
FromDateTime
This section contains reference documentation for the FromDateTime function.
Converts a formatted date-time string to milliseconds, based on the provided .
Signature
FromDateTime(dateTimeString, pattern)
FLOOR
This section contains reference documentation for the FLOOR function.
Rounded down to the nearest integer.
Signature
FLOOR(col1)
COVAR_POP
This section contains reference documentation for the COVAR_POP function.
Returns the population covariance between of 2 numerical columns.
Signatures
COVAR_POP(col1, col2) -> double
DISTINCTCOUNT
This section contains reference documentation for the DISTINCTCOUNT function.
Returns the count of distinct row values in a group
Signature
DISTINCTCOUNT(colName)
JSONFORMAT
This section contains reference documentation for the JSONFORMAT function.
Extracts the object value from jsonField based on 'jsonPath', the result type is inferred based on JSON value. This function can only be used in an .
Signature
JSONFORMAT(object)
DISTINCTCOUNTMV
This section contains reference documentation for the DISTINCTCOUNTMV function.
Returns the count of distinct row values in a group
Signature
DISTINCTCOUNTMV(colName)
MINMAXRANGEMV
This section contains reference documentation for the MINMAXRANGEMV function.
Returns the max - min value in a group
Signature
MINMAXRANGEMV(colName)
ltrim
This section contains reference documentation for the ltrim function.
trim spaces from left side of the string
Signature
LTRIM(col)
DISTINCTCOUNTBITMAPMV
This section contains reference documentation for the DISTINCTCOUNTHLLMV function.
Returns an approximate distinct count using HyperLogLog in a group.
Signature
DISTINCTCOUNTHLLMV(colName)
MINMV
This section contains reference documentation for the MINMV function.
Get the minimum value in a group
Signature
MINMV(colName)
MAXMV
This section contains reference documentation for the MAXMV function.
Get the maximum value in a group
Signature
MAXMV(colName)
lpad
This section contains reference documentation for the LPAD function.
string padded from the left side with pad to reach final size
Signature
LPAD(col, size, pad)
FromEpoch
This section contains reference documentation for the fromEpoch functions.
Convert epoch to epoch milliseconds. The following time units are supported:
SECONDS
MINUTES
mult
This section contains reference documentation for the MULT function.
Product of at least two values
Signature
MULT(col1, col2, col3...)
mode
This section contains reference documentation for the mode function.
Get the most frequent value in a group. When multiple modes are present it gives the minimum of all the modes. This behavior can be overridden to get the maximum or the average mode.
Signature
MODE(colName, [reducerType])
arrayConcatDouble
This section contains reference documentation for the arrayConcatDouble function.
Concatenates two arrays of doubles.
Signature
arrayConcatDouble('colName1', 'colName2')
max
This section contains reference documentation for the max function.
Putting exprmin/exprmax column inside order by clause (e.g. SELECT intCol, exprmin(longCol, doubleCol) FROM table GROUP BY intCol ORDER BY exprmin(longCol, doubleCol)) is not supported as semantically ordering multi-column multi-row exprmin/exprmax results doesn't make sense
Currently projecting MV bytes column doesn't work for now due to an issue
select homeRuns, numberOfGames, DIV(homeRuns, numberOfGames) AS total
from baseballStats
WHERE teamID = 'ML1'
AND yearID = 1956
AND playerName = 'Henry Louis'
select homeRuns, baseOnBalls, MULT(homeRuns, baseOnBalls) AS total
from baseballStats
WHERE teamID = 'ML1'
AND yearID = 1956
AND playerName = 'Henry Louis'
select DivAirportIDs,
arrayContainsInt(DivAirportIDs, 14683) AS containsValue
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
limit 5
SELECT EXPR_MAX(user_id, activity, last_activity_date, user_id)
FROM userEngagmentTable
SELECT user_region, EXPR_MAX(user_id, activity, last_activity_date, user_id),
EXPR_MIN(user_id, user_satisfaction)
FROM userEngagmentTable
GROUP BY user_region
select DivTailNums,
arrayDistinctString(DivTailNums) AS unique
from airlineStats
WHERE arraylength(DivTailNums) >= 2
limit 5
select DivAirportIDs,
arraySortInt(DivAirportIDs) AS sortedIds
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
limit 5
select DivTailNums,
arrayConcatString(DivTailNums, DivTailNums) AS concatIds
from airlineStats
WHERE arraylength(DivTailNums) >= 2
limit 5
select DivWheelsOffs,
DivWheelsOns,
arrayUnionInt(DivWheelsOffs, DivWheelsOns) AS unionIds
from airlineStats
WHERE arraylength(DivWheelsOffs) >= 2
limit 5
select FlightNum,
arrayReverseString(RandomAirports) AS reversedAirports,
RandomAirports
from airlineStats
WHERE arraylength(RandomAirports) BETWEEN 2 AND 4
limit 5
select RandomAirports,
arrayRemoveString(RandomAirports, 'SEA') AS value
from airlineStats
WHERE arraylength(RandomAirports) BETWEEN 2 AND 4
limit 5
select DISTINCT league AS value
from baseballStats
select DISTINCT(league) AS value
from baseballStats
SELECT AVG(DISTINCT AtBatting) AS VALUE
FROM baseballStats
select day(1639351800000, 'CET') AS day
FROM ignoreMe
select dayOfMonth(1639351800000) AS day
FROM ignoreMe
select dayOfMonth(1639351800000, 'CET') AS day
FROM ignoreMe
select DISTINCTCOUNTBITMAP(teamID) AS value
from baseballStats
select DISTINCTCOUNTBITMAPMV(DivTailNums) AS value
from airlineStats
where arraylength(DivTailNums) > 1
select DISTINCTCOUNTRAWHLL(teamID, 1) AS value
from baseballStats
select dayOfYear(1639351800000, 'CET') AS dayOfYear
FROM ignoreMe
select doy(1639351800000) AS dayOfYear
FROM ignoreMe
select doy(1639351800000, 'CET') AS dayOfYear
FROM ignoreMe
select DivAirportIDs,
arrayRemoveInt(DivAirportIDs, 12892) AS value
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
AND arrayContainsInt(DivAirportIDs, 12892) = 1
limit 5
select COUNTMV(DivTailNums) AS value
from airlineStats
where arraylength(DivTailNums) > 1
SELECT FromDateTime('2019-08-07', 'yyyy-MM-dd') AS epochMillis
FROM ignoreMe
SELECT FromDateTime(
'2019-08-07 3:12:13 PM',
'yyyy-MM-dd hh:mm:ss a'
) AS epochMillis
FROM ignoreMe
SELECT FromDateTime(
'2019-08-07T15:12:13',
'yyyy-MM-dd''T''HH:mm:ss'
) AS epochMillis
FROM ignoreMe
SELECT FromDateTime(
'2019-08-07T07:12:13-0800',
'yyyy-MM-dd''T''HH:mm:ssZ'
) AS epochMillis
FROM ignoreMe
This section contains reference documentation for the hour function.
Returns the hour of the day from the given epoch millis in UTC or specified timezone. The value ranges from 0 to 23.
Signature
hour(tsInMillis)
hour(tsInMillis, timeZoneId)
Usage Examples
hour
hour
Histogram
This section contains reference documentation for the HISTOGRAM function.
Returns the count of data points that fall within each bin as a vector. The bins are left-inclusive and right-exclusive, i.e. [a, b), except for the last one which is inclusive on both sides [a, b].
This section contains reference documentation for the arrayReverseInt function.
Reverses array of ints.
Signature
arrayReverseInt('colName')
Usage Examples
These examples are based on the .
DivAirportIDs
reversedIds
DISTINCTSUM
This section contains reference documentation for the DISTINCTSUM function.
Returns the sum of distinct row values in a group
Signature
DISTINCTSUM(colName) or sum(distinct col)
Usage Examples
These examples are based on the .
VALUE
VALUE
month
This section contains reference documentation for the month function.
Returns the month of the year from the given epoch millis in UTC or specified timezone. The value ranges from 1 to 12.
Signature
month(tsInMillis)
month(tsInMillis, timeZoneId)
Usage Examples
month
month
FIRSTWITHTIME
This section contains reference documentation for the firstwithtime function.
Returns the value of dataColumn with the smallest timeColumn value where:
timeColumn is used to define the time of dataColumn, which can be of type TIMESTAMP, INT, LONG
dataType specifies the type for dataColumn, which can be BOOLEAN, INT, LONG, FLOAT, DOUBLE, STRING
Signature
FIRSTWITHTIME(dataColumn, timeColumn, 'dataType')
Example
This example is based on the .
value
DISTINCTCOUNTHLL
This section contains reference documentation for the DISTINCTCOUNTHLL function.
Returns an approximate distinct count using HyperLogLog. It also takes an optional second argument to configure the log2m for the HyperLogLog.
For accurate distinct counting, see DISTINCTCOUNT.
Signature
DISTINCTCOUNTHLL(colName, log2m)
Usage Examples
These examples are based on the .
value
value
jsonextractkey
This section contains reference documentation for the JSONEXTRACTKEY function.
Extracts all matched JSON field keys based on 'jsonPath' into a STRING_ARRAY.
Signature
JSONEXTRACTKEY(jsonField, 'jsonPath')
Arguments
Description
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
Usage Examples
The examples in this section are based on the . In particular we'll be querying the row WHERE id = 7044874109.
id
repo
keys
DISTINCTCOUNTRAWHLLMV
This section contains reference documentation for the DISTINCTCOUNTRAWHLLMV function.
Returns HLL response serialized as string. The serialized HLL can be converted back into an HLL and then aggregated with other HLLs. A common use case may be to merge HLL responses from different Pinot tables, or to allow aggregation after client-side batching.
Signature
DISTINCTCOUNTRAWHLLMV(colName, log2m)
Usage Examples
These examples are based on the .
value
value
millisecond
This section contains reference documentation for the millisecond function.
Returns the millisecond of the second from the given epoch millis in UTC or specified timezone. The value ranges from 0 to 999.
Signature
millisecond(tsInMillis)
millisecond(tsInMillis, timeZoneId)
Usage Examples
millisecond
millisecond
isSubnetOf
This section contains reference documentation for the isSubnetOf function.
Takes 2 arguments of type STRING. The first argument is an ipPrefix, and the second argument is a single ipAddress. This function handles both IPv4 and IPv6 arguments.
Returns a boolean value checking if ipAddress is in the subnet of ipPrefix
Signatures
isSubnetOf(ipPrefix, ipAddress) -> boolean
Usage Examples
See the following sample queries where isSubnetOf is used in different parts of the query.
min
This section contains reference documentation for the min function.
Get the minimum value in a group
Signature
MIN(colName)
Usage Examples
These examples are based on the .
value
MD5
This section contains reference documentation for the MD5 function.
Return MD5 digest of binary column(bytes type) as hex string
Signature
MD5(bytesCol)
Usage Examples
These examples are based on the .
event_id
location
hash
The row returned will be different if you run this example as the data is ingested in real-time.
minute
This section contains reference documentation for the minute function.
Returns the minute of the hour from the given epoch millis in UTC or specified timezone. The value ranges from 0 to 59.
Signature
minute(tsInMillis)
minute(tsInMillis, timeZoneId)
Usage Examples
minute
minute
FrequentLongsSketch
This section contains reference documentation for the FREQUENTLONGSSKETCH function.
FREQUENTLONGSSKETCH is an estimation data-sketch function which can be used to estimate the frequency of an item. It is based on Apache Datasketches library and returns a serialized sketch object which can be merged with other sketches.
column (required): Name of the column to aggregate on. Needs to be a type which can be cast into 'LONG'.
maxMapSize: This value specifies the maximum physical length of the internal hash map. The maxMapSize must be a power of 2 and the default value is 256.
Usage Example
frequentlongssketch(AirlineID)
Which can be used, for example in Java as:
For more examples on the sketch API, refer to the Datasketches .
MOD
This section contains reference documentation for the MOD function.
Modulo of two values
Signature
MOD(col1, col2)
Usage Examples
value
value
caseWhen
This section contains reference documentation for the caseWhen function.
Returns values depending on boolean expressions. This function can only be used in an .
This section contains reference documentation for the JSONPATHDOUBLE function.
Extracts the Double value from jsonField based on 'jsonPath', use optional defaultValuefor null or parsing error. This function can only be used in an .
Signature
Base64
This section contains reference documentation for base64 encode and decode functions.
Encoding scheme follows
toBase64 returns Base64 encoded string of input binary data (bytes type).
fromBase64
FrequentStringsSketch
This section contains reference documentation for the FREQUENTSTRINGSSKETCH function.
FREQUENTSTRINGSSKETCH is an estimation data-sketch function which can be used to estimate the frequency of an item. It is based on and returns a serialized sketch object which can be merged with other sketches.
select millisecond(1639351800000) AS millisecond
FROM ignoreMe
0
select millisecond(1639351800000, 'America/St_Johns') AS millisecond
FROM ignoreMe
0
SELECT isSubnetOf('192.168.0.1/24', '192.168.0.225')
AS result
FROM myTable;
---> returns true
SELECT isSubnetOf('1.2.3.128/26', '1.2.5.1')
AS result
FROM myTable;
---> returns false
SELECT isSubnetOf('2001:4800:7825:103::/64', '2001:4800:7825:103::2050')
AS result
FROM myTable;
---> returns true
SELECT isSubnetOf('7890:db8:113::8a2e:370:7334/127', '7890:db8:113::8a2e:370:7336')
AS result
FROM myTable;
---> returns false
returns binary data (represented as a Hex string) from Base64-encoded string.
Signature
toBase64(bytesCol)
fromBase64(stringCol)
Usage Examples
For better readability, the following examples converts string hello! into BYTES using toUtf8 function and converts the decoded BYTES into string using fromUtf8.
encoded
aGVsbG8h
decoded
hello!
Note that without UTF8 string conversion, returned BYTES will be represented as a Hex string following Pinot's BYTES column representation. See the example below.
decoded
68656c6c6f21
Note that the following query will throw compilation error as string is not a valid input type for toBase64.
column (required): Name of the column to aggregate on. Needs to be a type which can be cast into 'STRING'.
maxMapSize: This value specifies the maximum physical length of the internal hash map. The maxMapSize must be a power of 2 and the default value is 256.
Usage Example
frequentstringssketch(AirlineID)
BAEKCAUAAAAOAA...
Which can be used, for example in Java as:
For more examples on the sketch API, refer to the Datasketches documentation.
select FlightNum,
arraySliceString(RandomAirports, 0, 2) AS airports,
RandomAirports
from airlineStats
WHERE arraylength(RandomAirports) BETWEEN 2 AND 4
limit 5
select dayOfWeek(1639351800000, 'CET') AS dayOfWeek
FROM ignoreMe
select dow(1639351800000) AS dayOfWeek
FROM ignoreMe
select dow(1639351800000, 'CET') AS dayOfWeek
FROM ignoreMe
select FlightNum,
arraySortString(RandomAirports) AS sortedAirports,
RandomAirports
from airlineStats
WHERE arraylength(RandomAirports) BETWEEN 2 AND 4
limit 5
select DivTailNums,
DivAirports,
arrayUnionString(DivTailNums, DivAirports) AS unionIds
from airlineStats
WHERE arraylength(DivTailNums) >= 2
limit 5
select ARRAYLENGTH(RandomAirports) AS length, count(*)
from airlineStats
GROUP BY length
ORDER BY count(*) DESC
LIMIT 5
select FromEpochSecondsBucket(1613472303, 1) AS bucket
FROM ignoreMe
select FromEpochSecondsBucket(1613472303, 2) AS bucket
FROM ignoreMe
select FromEpochMinutesBucket(2689120, 10) AS bucket
FROM ignoreMe
select FromEpochHoursBucket(89637, 5) AS bucket
FROM ignoreMe
select FromEpochDaysBucket(1867, 10) AS bucket
FROM ignoreMe
SELECT HISTOGRAM(numberOfGames, 0, 200, 10) AS histogram
FROM baseballStats
select HISTOGRAM(AtBatting, Array['-Infinity', 1, 10, 50, 100, 500, 1000]) AS histogram
from baseballStats
select DivAirportIDs,
arrayReverseInt(DivAirportIDs) AS reversedIds
from airlineStats
WHERE arraylength(DivAirportIDs) >= 2
limit 5
select FIRSTWITHTIME(group_name, __metadata$recordTimestamp, 'STRING')
from meetupRsvp
select DISTINCTCOUNTHLL(teamID, 12) AS value
from baseballStats
select id, repo, JSONEXTRACTKEY(repo, '$.*') AS keys
from githubEvents
WHERE id = 7044874109
select DISTINCTCOUNTRAWHLLMV(DivAirports, 1) AS value
from airlineStats
where arraylength(DivAirports) > 1
SELECT count(*)
FROM myTable
WHERE isSubnetOf('192.168.0.1/24', ipAddressCol);
SELECT count(*)
FROM myTable
WHERE isSubnetOf('192.168.0.1/24', ipAddressCol)
OR isSubnetOf(ipPrefixCol, '7890:db8:113::8a2e:370:7336');
SELECT
CASE
WHEN isSubnetOf('105.25.245.115/27', srcIPAddress) THEN 'case1'
WHEN isSubnetOf('105.25.245.115/27', dstIPAddress) THEN 'case2'
ELSE 'case3'
END AS differentFlow
FROM myTable;
select FREQUENTLONGSSKETCH(AirlineID, 16) from airlineStats
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the Jayway JsonPath Evaluator Tool to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
Expression
Value
JSONPATHDOUBLE(data, '$.age')
24.0
This function can be used in the table config to extract the age property into the age column, as described below:
'jsonPath'and`` ``'results_type'are literals. Pinot uses single quotes to distinguish them from identifiers.
Usage Examples
The examples in this section are based on the . In particular we'll be querying the row WHERE id = 7044874109:
repo
The following examples show how to use the JSONEXTRACTSCALAR function:
id
name
id
name
DISTINCTCOUNTTHETASKETCH
This section contains reference documentation for the DISTINCTCOUNTTHETASKETCH function.
The Theta Sketch framework enables set operations over a stream of data, and can also be used for cardinality estimation. Pinot leverages the Sketch Class and its extensions from the library org.apache.datasketches:datasketches-java:4.2.0 to perform distinct counting as well as evaluating set operations.
Signature
distinctCountThetaSketch(<thetaSketchColumn>, <thetaSketchParams>, predicate1, predicate2..., postAggregationExpressionToEvaluate) -> Long
thetaSketchColumn (required): Name of the column to aggregate on.
thetaSketchParams (required): Parameters for constructing the intermediate theta-sketches.
Usage Examples
These examples are based on the .
value
value
We can also provide predicates and a post aggregation expression to compute more complicated cardinalities. For example, we could can find the intersection of the following queries:
yearID
yearID
(the yearId 1986 is the only one in common)
By running the following query:
value
JSONPATHSTRING
This section contains reference documentation for the JSONPATHSTRING function.
Extracts the String value from jsonField based on 'jsonPath', use optional defaultValuefor null or parsing error. This function can only be used in an ingestion transformation function.
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
Expression
Value
This function can be used in the to extract the age property into the age column, as described below:
DISTINCTCOUNTRAWTHETASKETCH
This section contains reference documentation for the DISTINCTCOUNTRAWTHETASKETCH function.
The Theta Sketch framework enables set operations over a stream of data, and can also be used for cardinality estimation. Pinot leverages the Sketch Class and its extensions from the library org.apache.datasketches:datasketches-java:4.2.0 to perform distinct counting as well as evaluating set operations.
Currently, the only supported parameter is nominalEntries (defaults to 4096).
predicates (optional)_: _ These are individual predicates of form lhs <op> rhs which are applied on rows selected by the where clause. During intermediate sketch aggregation, sketches from the thetaSketchColumn that satisfies these predicates are unionized individually. For example, all filtered rows that match country=USA are unionized into a single sketch. Complex predicates that are created by combining (AND/OR) of individual predicates is supported.
postAggregationExpressionToEvaluate (required): The set operation to perform on the individual intermediate sketches for each of the predicates. Currently supported operations are SET_DIFF, SET_UNION, SET_INTERSECT , where DIFF requires two arguments and the UNION/INTERSECT allow more than two arguments.
Currently, the only supported parameter is nominalEntries (defaults to 4096).
predicates (optional)_: _ These are individual predicates of form lhs <op> rhs which are applied on rows selected by the where clause. During intermediate sketch aggregation, sketches from the thetaSketchColumn that satisfies these predicates are unionized individually. For example, all filtered rows that match country=USA are unionized into a single sketch. Complex predicates that are created by combining (AND/OR) of individual predicates is supported.
postAggregationExpressionToEvaluate (required): The set operation to perform on the individual intermediate sketches for each of the predicates. Currently supported operations are SET_DIFF, SET_UNION, SET_INTERSECT , where DIFF requires two arguments and the UNION/INTERSECT allow more than two arguments.
select id, jsonextractscalar(repo, '$.foo', 'STRING', 'dummyValue') AS name
from githubEvents
WHERE id = 7044874109
select distinctCountThetaSketch(teamID) AS value
from baseballStats
select distinctCountThetaSketch(teamID, 'nominalEntries=10') AS value
from baseballStats
select yearID
from baseballStats
where teamID = 'SFN' AND numberOfGames = 28 AND homeRuns = 1
select yearID
from baseballStats
where teamID = 'CHN' AND numberOfGames = 28 AND homeRuns = 1
select distinctCountThetaSketch(
yearID,
'nominalEntries=4096',
'teamID = ''SFN'' AND numberOfGames=28 AND homeRuns=1',
'teamID = ''CHN'' AND numberOfGames=28 AND homeRuns=1',
'SET_INTERSECT($1, $2)'
) AS value
from baseballStats
select distinctCountRawThetaSketch(teamID) AS value
from baseballStats
select distinctCountRawThetaSketch(teamID, 'nominalEntries=10') AS value
from baseballStats
select distinctCountRawThetaSketch(
yearID,
'nominalEntries=4096',
'teamID = ''SFN'' AND numberOfGames=28 AND homeRuns=1',
'teamID = ''CHN'' AND numberOfGames=28 AND homeRuns=1',
'SET_INTERSECT($1, $2)'
) AS value
from baseballStats
Arguments
Description
jsonField
An Identifier/Expression contains JSON documents.
'jsonPath'
Follows to read values from JSON documents.
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the Jayway JsonPath Evaluator Tool to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
This function can be used in the table config to extract the name, score, and second value of homework_grades into their respective columns , as described below:
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the Jayway JsonPath Evaluator Tool to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
Expression
Value
JSONPATHLONG(data, '$.age')
24
This function can be used in the table config to extract the age property into the age column, as described below:
This section contains reference documentation for the JSONPATH function.
Extracts the object value from jsonField based on 'jsonPath', the result type is inferred based on JSON value. This function can only be used in an ingestion transformation function.
Signature
JSONPATH(jsonField, 'jsonPath')
Arguments
Description
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
Expression
Value
This function can be used in the to extract the name property into the name column and age property into the age column, as described below:
inputFormat and outputFormat are defined using the following structure:
<time size>:<time unit>:<time format>:<pattern>
where:
time size - size of the time unit eg: 1, 10
time unit - DAYS, HOURS, MINUTES, SECONDS, MILLISECONDS
granularity is specified in the format <time size>:<time unit>.
Usage Examples
These examples are based on the .
created_at_timestamp from milliseconds since epoch to days since epoch, bucketed to 1 day granularity:
id
created_at_timestamp
timeInMs
convertedTime
created_at_timestamp bucketed to 15 minutes granularity:
id
created_at_timestamp
timeInMs
convertedTime
created_at_timestamp to format yyyy-MM-dd, bucketed to 1 days granularity:
id
created_at_timestamp
timeInMs
convertedTime
created_at_timestamp to format yyyy-MM-dd HH:mm, in timezone Pacific/Kiritimati:
id
created_at_timestamp
timeInMs
convertedTime
created_at_timestamp to format yyyy-MM-dd, in timezone Pacific/Kiritimati and bucketed to 1 day granularity:
id
created_at_timestamp
timeInMs
convertedTime
DATETRUNC
This section contains reference documentation for the DATETRUNC function.
(Presto) SQL compatible date truncation, equivalent to the Presto function date_trunc.
Converts the value into a specified output granularity seconds since UTC epoch that is bucketed on a unit in a specified timezone.
Signature
JSONPATHARRAYDEFAULTEMPTY
This section contains reference documentation for the JSONPATHARRAYDEFAULTEMPTY function.
Extracts an array from jsonField based on 'jsonPath', the result type is inferred based on JSON value. Returns empty array for null or parsing error. This function can only be used in an .
Signature
,
MICROSECONDS
,
NANOSECONDS
time format
EPOCH
SIMPLE_DATE_FORMAT pattern - defined in case of SIMPLE_DATE_FORMAT e.g. yyyy-MM-dd. A specific timezone can be passed using tz(timezone). Timezone can be long or short string format timezone. e.g. Asia/Kolkata or PDT
inputTimeUnitStr and outputTimeUnitStr support the following values:
NANOSECONDS
MICROSECONDS
MILLISECONDS
SECONDS
MINUTES
HOURS
DAYS
Usage Examples
Truncates an epoch in milliseconds at WEEK (where a Week starts at Monday UTC midnight):
or
ts
1639353600000
Truncates an epoch in milliseconds at WEEK (where a Week starts at Monday UTC midnight) in the UTC time zone, returning a result in epoch in seconds in UTC timezone:
ts
1639353600
Truncates an epoch in milliseconds at WEEK (where a Week starts at Monday UTC midnight) in the CET time zone, returning a result in epoch in seconds in UTC timezone:
ts
1639350000
Truncates an epoch in milliseconds at QUARTER in the Los Angeles time zone (where a Quarter begins on Jan 1st, April 1st, July 1st, October 1st in Los Angeles timezone), returning a result in hours since UTC epoch:
ts
453631
select id,
created_at_timestamp,
cast(created_at_timestamp AS long) AS timeInMs,
DATETIMECONVERT(
created_at_timestamp,
'1:MILLISECONDS:EPOCH',
'1:DAYS:EPOCH',
'1:DAYS'
) AS convertedTime
from githubEvents
WHERE id = 7044874134
select id,
created_at_timestamp,
cast(created_at_timestamp AS long) AS timeInMs,
DATETIMECONVERT(
created_at_timestamp,
'1:MILLISECONDS:EPOCH',
'1:MILLISECONDS:EPOCH',
'15:MINUTES'
) AS convertedTime
from githubEvents
WHERE id = 7044874134
select id,
created_at_timestamp,
cast(created_at_timestamp AS long) AS timeInMs,
DATETIMECONVERT(
created_at_timestamp,
'1:MILLISECONDS:EPOCH',
'1:DAYS:SIMPLE_DATE_FORMAT:yyyy-MM-dd',
'1:DAYS'
) AS convertedTime
from githubEvents
WHERE id = 7044874134
select id,
created_at_timestamp,
cast(created_at_timestamp AS long) AS timeInMs,
DATETIMECONVERT(
created_at_timestamp,
'1:MILLISECONDS:EPOCH',
'1:MILLISECONDS:SIMPLE_DATE_FORMAT:yyyy-MM-dd HH:mm tz(Pacific/Kiritimati)',
'1:MILLISECONDS'
) AS convertedTime
from githubEvents
WHERE id = 7044874134
select id,
created_at_timestamp,
cast(created_at_timestamp AS long) AS timeInMs,
DATETIMECONVERT(
created_at_timestamp,
'1:MILLISECONDS:EPOCH',
'1:MILLISECONDS:SIMPLE_DATE_FORMAT:yyyy-MM-dd HH:mm tz(Pacific/Kiritimati)',
'1:DAYS'
) AS convertedTime
from githubEvents
WHERE id = 7044874134
select dateTrunc('week', 1639480981746) AS ts
FROM ignoreMe
select dateTrunc('week', 1639480981746, 'MILLISECONDS') AS ts
FROM ignoreMe
select dateTrunc(
'week',
1639480981746,
'MILLISECONDS',
'UTC',
'SECONDS'
) AS ts
FROM ignoreMe
select dateTrunc(
'week',
1639480981746,
'MILLISECONDS',
'CET',
'SECONDS'
) AS ts
FROM ignoreMe
select dateTrunc(
'quarter',
1639480981746,
'MILLISECONDS',
'America/Los_Angeles',
'HOURS'
) AS ts
FROM ignoreMe
JSONPATHARRAYDEFAULTEMPTY(jsonField, 'jsonPath')
Arguments
Description
jsonField
An Identifier/Expression contains JSON documents.
'jsonPath'
Follows to read values from JSON documents.
'jsonPath'` is a literal. Pinot uses single quotes to distinguish them from identifiers.
You can use the Jayway JsonPath Evaluator Tool to test JSON expressions before you import any data.
Usage Examples
The usage examples are based on extracting fields from the following JSON document:
This function can be used in the table config to extract the name, score, and second value of homework_grades into their respective columns , as described below:
(required) These are individual predicates representing funnel steps which are applied on rows selected by the where clause. Distinct values from the correlation_column that satisfy these predicates are counted per step. For example, all filtered rows that match url = '/checkout' are unionized into a set. Sets are intersected with the sets resulted from the preceding steps, each step retaining only individuals present in previous steps. Finally, unique counts are returned for each step in the funnel.
CORRELATED_BY
correlation_column
(required) Column to leverage for funnel correlation, distinct values from this column are counted per step during aggregation. Only dictionary-encoded columns are supported.
SETTINGS
settings 1...n
(optional) Settings to select and configure a funnel counting strategy:
bitmap (default): This strategy is accurate for INT column, but approximate for other cases where hash codes are used in distinct counting and there may be hash collisions. For accurate distinct counting on all column types, use 'set' instead. See also .
set: This strategy uses hash sets. Use with care, unbounded memory cost. See also .
theta_sketch: This strategy leverages framework to provide an approximate funnel count with a small memory footprint. See also .
Usage Examples
Many datasets are time series in nature, tracking events of an entity over time. An example of such a dataset could be a user analytics activity log from a commerce web application.
Example
user_id
event_time
url
U1
2021-10-01 09:01:00.000
/product/listing
U2
2021-10-01 09:17:00.000
/product/search
U1
2021-10-01 09:33:00.000
/product/details
Funnel
We want to analyse the following checkout funnel:
/cart/add
/checkout/start
/checkout/confirmation
Counts
We want to answer the following questions about the above funnel:
How many users entered the top of the funnel?
How many of these users proceeded to the second step?
How many users reached the bottom of the funnel after completing all steps?
Query
counts
3, 2, 2
Notes
Notice that although U1 user added to cart twice, it still counted as one conversion in the first step, as we report on unique counts rather than total events. Also notice that although U2 events were logged out of order, we still counted the user as converted.
Equivalence
The above query is equivalent to the below presto SQL query:
Settings
For a large dataset we could use for example a theta_sketch strategy, or furthermore, partition the data by user_id and leverage a partitioned strategy. It is also important to filter in the where clause so to aggregate only necessary rows.
counts
3, 2, 2
Another Example
We now want to learn how many users checkout after a text search; as opposed to other entry points such as browsing a product category listing. We want to then analyse the following funnel:
/product/search
/cart/add
/checkout/start
/checkout/confirmation
Query
counts
2, 2, 1, 1
Notes
Notice that U1 is not counted in this funnel, as the user did not perform any product search. Both U2 and U3 entered the top of the funnel and performed the second step, but only U2 converted to the bottom of the funnel.
select
ARRAY[
count_if(steps[1]),
count_if(steps[1] and steps[2]),
count_if(steps[1] and steps[2] and steps[3])
] as counts
from (
select
ARRAY[
bool_or(url = '/cart/add'),
bool_or(url = '/checkout/start'),
bool_or(url = '/checkout/confirmation')
] as steps
from user_log
group by user_id
)
select
FUNNEL_COUNT(
STEPS(
url = '/cart/add',
url = '/checkout/start',
url = '/checkout/confirmation'),
CORRELATED_BY(user_id)
) AS counts
from user_log
select
FUNNEL_COUNT(
STEPS(
url = '/cart/add',
url = '/checkout/start',
url = '/checkout/confirmation'),
CORRELATED_BY(user_id),
SETTINGS('theta_sketch', 'nominalEntries=4096')
) AS counts
from user_log
where url in ('/cart/add', '/checkout/start', '/checkout/confirmation')
nominalEntries: theta-sketch strategy parameter (defaults to 4096). Can only be used in conjunction with theta_sketch setting.
partitioned: This strategy counts funnel steps per segment, then sums up step counts across segments. Correlation column should be configured as partition column for this strategy. See also .
sorted: This strategy counts funnel steps per segment with zero memory footprint. Correlation column should be configured as sort column for this strategy. Can only be used in conjunction with partitioned setting.