Title: | Interface to 'Kusto'/'Azure Data Explorer' |
---|---|
Description: | An interface to 'Azure Data Explorer', also known as 'Kusto', a fast, distributed data exploration service from Microsoft: <https://azure.microsoft.com/en-us/products/data-explorer/>. Includes 'DBI' and 'dplyr' interfaces, with the latter modelled after the 'dbplyr' package, whereby queries are translated from R into the native 'KQL' query language and executed lazily. On the admin side, the package extends the object framework provided by 'AzureRMR' to support creation and deletion of databases, and management of database principals. Part of the 'AzureR' family of packages. |
Authors: | Hong Ooi [aut], Alex Kyllo [aut, cre], dbplyr development team [cph] (Original framework for dplyr/database interface), Microsoft [cph] |
Maintainer: | Alex Kyllo <[email protected]> |
License: | MIT + file LICENSE |
Version: | 1.1.3.9000 |
Built: | 2024-08-26 04:03:13 UTC |
Source: | https://github.com/azure/azurekusto |
Append a join operation to the tbl_kusto object's ops list
add_op_join( type, x, y, by = NULL, suffix = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL )
add_op_join( type, x, y, by = NULL, suffix = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL )
type |
The name of the join type, one of: inner_join, left_join, right_join, full_join, semi_join, anti_join |
x |
The "left" tbl |
y |
The "right" tbl |
by |
A vector of column names; keys by which tbl x and tbl y will be joined |
suffix |
A vector of strings that will be appended to the names of non-join key columns that exist in both tbl x and tbl y to distinguish them by source tbl. |
.strategy |
A strategy hint to provide to Kusto. |
.shufflekeys |
A character vector of column names to shuffle on, if |
.num_partitions |
The number of partitions for a shuffle query. |
.remote |
A strategy hint to provide to Kusto for cross-cluster joins. |
Append a set operation to the tbl_kusto object's ops list
add_op_set_op(x, y, type)
add_op_set_op(x, y, type)
x |
The "left" tbl |
y |
The "right" tbl |
type |
The type of set operation to perform, currently only supports union_all |
Append an operation representing a single-table verb to the tbl_kusto object's ops list
add_op_single(name, .data, dots = list(), args = list())
add_op_single(name, .data, dots = list(), args = list())
name |
The name of the operation, e.g. 'select', 'filter' |
.data |
The tbl_kusto object to append the operation to |
dots |
The expressions passed as arguments to the operation verb |
args |
Other non-expression arguments passed to the operation verb |
Class representing a Kusto cluster, exposing methods for working with it.
The following methods are available, in addition to those provided by the AzureRMR::az_resource class:
new(...)
: Initialize a new storage object. See 'Initialization'.
start()
: Start the cluster.
stop()
: Stop the cluster.
create_database(...)
: Create a new Kusto database. See Databases
below.
get_database(database))
: Get an existing database.
delete_database(database, confirm=TRUE)
: Delete a database, by default asking for confirmation first.
list_databases()
: List all databases in this cluster.
get_default_tenant()
: Retrieve the default tenant to authenticate with this cluster.
get_query_token(tenant, ...)
: Obtain an authentication token from Azure Active Directory for this cluster's query endpoint. Accepts further arguments that will be passed to get_kusto_token.
get_ingestion_token(tenant, ...)
: Obtain an authentication token for this cluster's ingestion endpoint. Accepts further arguments that will be passed to get_kusto_token.
Initializing a new object of this class can either retrieve an existing Kusto cluster, or create a new cluster on the host. Generally, the best way to initialize an object is via the get_kusto_cluster
and create_kusto_cluster
methods of the az_resource_group class, which handle the details automatically.
A Kusto cluster can have several databases, which are represented in AzureKusto via az_kusto_database R6 objects. The az_kusto
class provides the create_database
, get_database
, delete_database
and list_databases
methods for creating, deleting and retrieving databases. It's recommended to use these methods rather than calling az_kusto_database$new()
directly.
create_database
takes the following arguments. It returns an object of class az_kusto_database
database
: The name of the database to create.
retention_period
: The retention period of the database, after which data will be soft-deleted.
cache_period
: The cache period of the database, the length of time for which queries will be cached.
get_database
takes a single argument database
, the name of the database to retrieve, and returns an object of class az_kusto_database
. delete_database
takes the name of the database to delete and returns NULL on a successful deletion. list_databases
takes no arguments and returns a list of az_kusto_database
objects, one for each database in the cluster.
az_kusto_database, kusto_database_endpoint, create_kusto_cluster, get_kusto_cluster, delete_kusto_cluster, get_kusto_token
Kusto/Azure Data Explorer documentation,
## Not run: # recommended way of retrieving a resource: via a resource group object kus <- resgroup$get_kusto_cluster("mykusto") # list databases kust$list_databases() # create a new database with a retention period of 6 months kust$create_database("newdb", retention_period=180) # get the default authentication tenant kus$get_default_tenant() # generate an authentication token kust$get_aad_token() ## End(Not run)
## Not run: # recommended way of retrieving a resource: via a resource group object kus <- resgroup$get_kusto_cluster("mykusto") # list databases kust$list_databases() # create a new database with a retention period of 6 months kust$create_database("newdb", retention_period=180) # get the default authentication tenant kus$get_default_tenant() # generate an authentication token kust$get_aad_token() ## End(Not run)
Class representing a Kusto database, exposing methods for working with it.
The following methods are available, in addition to those provided by the AzureRMR::az_resource class:
new(...)
: Initialize a new storage object. See 'Initialization'.
add_principals(...)
: Add new database principals. See Principals
below.
remove_principals(...)
: Remove database principals.
list_principals()
: Retrieve all database principals, as a data frame.
get_query_endpoint()
: Get a query endpoint object for interacting with the database.
get_ingestion_endpoint()
: Get an ingestion endpoint object for interacting with the database.
Initializing a new object of this class can either retrieve an existing Kusto database, or create a new database on the server. Generally, the best way to initialize an object is via the get_database
, list_databases()
and create_database
methods of the az_kusto class, which handle the details automatically.
This class provides methods for managing the principals of a database.
add_principal
takes the following arguments. It returns a data frame with one row per principal, containing the details for each principal.
name
: The name of the principal to create.
role
: The role of the principal, for example "Admin" or "User".
type
: The type of principal, either "User" or "App".
fqn
: The fully qualified name of the principal, for example "aaduser=username@mydomain" for an Azure Active Directory account. If supplied, the other details will be obtained from this.
email
: For a user principal, the email address.
app_id
: For an application principal, the ID.
remove_principal
removes a principal. It takes the same arguments as add_principal
; if the supplied details do not match the actual details for the principal, it is not removed.
az_kusto, kusto_database_endpoint, create_database, get_database, delete_database
Kusto/Azure Data Explorer documentation,
## Not run: # recommended way of retrieving a resource: via a resource group object db <- resgroup$ get_kusto_cluster("mykusto")$ get_database("mydatabase") # list principals db$list_principals() # add a new principal db$add_principal("New User", role="User", fqn="aaduser=username@mydomain") # get the endpoint db$get_database_endpoint(use_integer64=FALSE) ## End(Not run)
## Not run: # recommended way of retrieving a resource: via a resource group object db <- resgroup$ get_kusto_cluster("mykusto")$ get_database("mydatabase") # list principals db$list_principals() # add a new principal db$add_principal("New User", role="User", fqn="aaduser=username@mydomain") # get the endpoint db$get_database_endpoint(use_integer64=FALSE) ## End(Not run)
Functions to connect to a Kusto cluster.
AzureKusto() ## S4 method for signature 'AzureKustoDriver' dbConnect(drv, ..., bigint = c("numeric", "integer64")) ## S4 method for signature 'AzureKustoDriver' dbCanConnect(drv, ...) ## S4 method for signature 'AzureKustoDriver' dbDisconnect(conn, ...)
AzureKusto() ## S4 method for signature 'AzureKustoDriver' dbConnect(drv, ..., bigint = c("numeric", "integer64")) ## S4 method for signature 'AzureKustoDriver' dbCanConnect(drv, ...) ## S4 method for signature 'AzureKustoDriver' dbDisconnect(conn, ...)
drv |
An AzureKusto DBI driver object, instantiated with |
... |
Authentication arguments supplied to |
bigint |
How to treat Kusto long integer columns. By default, they will be converted to R numeric variables. If this is "integer64", they will be converted to |
conn |
For |
Kusto is connectionless, so dbConnect
simply wraps a database endpoint object, generated with kusto_database_endpoint(...)
. The endpoint itself can be accessed via the @endpoint
slot. Similarly, dbDisconnect
always returns TRUE.
dbCanConnect
attempts to detect whether querying the database with the given information and credentials will be successful. The result may not be accurate; essentially all it does is check that its arguments are valid Kusto properties. Ultimately the best way to tell if querying will work is to try it.
For dbConnect
, an object of class AzureKustoConnection.
For dbCanConnect
, TRUE if authenticating with the Kusto server succeeded with the given arguments, and FALSE otherwise.
For dbDisconnect
, always TRUE, invisibly.
kusto-DBI, dbReadTable, dbWriteTable, dbGetQuery, dbSendStatement, kusto_database_endpoint
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.westus.kusto.windows.net", database="database", tenantid="contoso") DBI::dbDisconnect(db) # no authentication credentials: returns FALSE DBI::dbCanConnect(AzureKusto(), server="https://mycluster.westus.kusto.windows.net") ## End(Not run)
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.westus.kusto.windows.net", database="database", tenantid="contoso") DBI::dbDisconnect(db) # no authentication credentials: returns FALSE DBI::dbCanConnect(AzureKusto(), server="https://mycluster.westus.kusto.windows.net") ## End(Not run)
Aggregation function translations
base_agg
base_agg
An object of class environment
of length 7.
Scalar operator translations (infix and prefix)
base_scalar
base_scalar
An object of class environment
of length 79.
Window function translations
base_window
base_window
An object of class environment
of length 1.
Build a KQL string.
build_kql(..., .env = parent.frame())
build_kql(..., .env = parent.frame())
... |
input to convert to KQL. Use |
.env |
the environment in which to evaluate the arguments. Should not be needed in typical use. |
Compile the preceding dplyr operations into a kusto query, execute it on the remote server, and return the result as a tibble.
## S3 method for class 'tbl_kusto' collect(x, ...)
## S3 method for class 'tbl_kusto' collect(x, ...)
x |
An instance of class tbl_kusto representing a Kusto table |
... |
needed for agreement with generic. Not otherwise used. |
Execute the query, store the results in a table, and return a reference to the new table
## S3 method for class 'tbl_kusto' compute(x, ..., name = generate_table_name())
## S3 method for class 'tbl_kusto' compute(x, ..., name = generate_table_name())
x |
An instance of class tbl_kusto representing a Kusto table |
... |
other parameters passed to the query |
name |
The name for the Kusto table to be created. If name is omitted, the table will be named Rtbl_ + 8 random lowercase letters |
This function uploads a local data frame into a remote data source, creating the table definition as needed. If the table exists, it will append the data to the existing table. If not, it will create a new table.
## S3 method for class 'kusto_database_endpoint' copy_to( dest, df, name = deparse(substitute(df)), overwrite = FALSE, method = "inline", ... )
## S3 method for class 'kusto_database_endpoint' copy_to( dest, df, name = deparse(substitute(df)), overwrite = FALSE, method = "inline", ... )
dest |
remote data source |
df |
local data frame |
name |
Name for new remote table |
overwrite |
If |
method |
For local ingestion, the method to use. "inline", "streaming", or "indirect". |
... |
other parameters passed to the query |
collect()
for the opposite action; downloading remote data into a local tbl.
Method for the AzureRMR::az_resource_group class.
create_kusto_cluster(name, location, node_size="D14_v2", ...)
name
: The name of the cluster.
location
: The location/region in which to create the account. Defaults to the resource group location.
node_size
: The capacity of the nodes in each of the cluster. Defaults to "D14_v2", which should be available in all regions. The availability of other sizes depends on the region the cluster is created in.
... Other named arguments to pass to the az_kusto initialization function.
This method deploys a new Kusto cluster resource, with parameters given by the arguments.
An object of class az_kusto
representing the created cluster.
get_kusto_cluster, delete_kusto_cluster, az_kusto
Kusto/Azure Data Explorer documentation
## Not run: rg <- AzureRMR::get_azure_login("myaadtenant")$ get_subscription("subscription_id")$ get_resource_group("rgname") # create a new Kusto cluster rg$create_kusto_cluster("mykusto", node_size="L16") ## End(Not run)
## Not run: rg <- AzureRMR::get_azure_login("myaadtenant")$ get_subscription("subscription_id")$ get_resource_group("rgname") # create a new Kusto cluster rg$create_kusto_cluster("mykusto", node_size="L16") ## End(Not run)
DBI methods for Kusto queries and commands
## S4 method for signature 'AzureKustoConnection,character' dbGetQuery(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection' dbSendQuery(conn, statement, ...) ## S4 method for signature 'AzureKustoResult' dbFetch(res, n = -1, ...) ## S4 method for signature 'AzureKustoConnection,character' dbSendStatement(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection,character' dbExecute(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection,character' dbListFields(conn, name, ...) ## S4 method for signature 'AzureKustoResult' dbColumnInfo(res, ...)
## S4 method for signature 'AzureKustoConnection,character' dbGetQuery(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection' dbSendQuery(conn, statement, ...) ## S4 method for signature 'AzureKustoResult' dbFetch(res, n = -1, ...) ## S4 method for signature 'AzureKustoConnection,character' dbSendStatement(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection,character' dbExecute(conn, statement, ...) ## S4 method for signature 'AzureKustoConnection,character' dbListFields(conn, name, ...) ## S4 method for signature 'AzureKustoResult' dbColumnInfo(res, ...)
conn |
An AzureKustoConnection object. |
statement |
A string containing a Kusto query or control command. |
... |
Further arguments passed to |
res |
An AzureKustoResult resultset object |
n |
The number of rows to return. Not used. |
name |
For |
These are the basic DBI functions to query the database. Note that Kusto only supports synchronous queries and commands; in particular, dbSendQuery
and dbSendStatement
will wait for the query or statement to complete, rather than returning immediately.
dbSendStatement
and dbExecute
are meant for running Kusto control commands, and will throw an error if passed a regular query. dbExecute
also returns the entire result of running the command, rather than simply a row count.
dbConnect, dbReadTable, dbWriteTable, run_query
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.location.kusto.windows.net", database="database"...) DBI::dbGetQuery(db, "iris | count") DBI::dbListFields(db, "iris") # does the same thing as dbGetQuery, but returns an AzureKustoResult object res <- DBI::dbSendQuery(db, "iris | count") DBI::dbFetch(res) DBI::dbColumnInfo(res) DBI::dbExecute(db, ".show tables") # does the same thing as dbExecute, but returns an AzureKustoResult object res <- DBI::dbSendStatement(db, ".show tables") DBI::dbFetch(res) ## End(Not run)
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.location.kusto.windows.net", database="database"...) DBI::dbGetQuery(db, "iris | count") DBI::dbListFields(db, "iris") # does the same thing as dbGetQuery, but returns an AzureKustoResult object res <- DBI::dbSendQuery(db, "iris | count") DBI::dbFetch(res) DBI::dbColumnInfo(res) DBI::dbExecute(db, ".show tables") # does the same thing as dbExecute, but returns an AzureKustoResult object res <- DBI::dbSendStatement(db, ".show tables") DBI::dbFetch(res) ## End(Not run)
DBI methods for Kusto table management
## S4 method for signature 'AzureKustoConnection,character' dbReadTable(conn, name, ...) ## S4 method for signature 'AzureKustoConnection,ANY' dbWriteTable(conn, name, value, method, ...) ## S4 method for signature 'AzureKustoConnection' dbCreateTable(conn, name, fields, ..., row.names = NULL, temporary = FALSE) ## S4 method for signature 'AzureKustoConnection,ANY' dbRemoveTable(conn, name, ...) ## S4 method for signature 'AzureKustoConnection' dbListTables(conn, ...) ## S4 method for signature 'AzureKustoConnection,ANY' dbExistsTable(conn, name, ...)
## S4 method for signature 'AzureKustoConnection,character' dbReadTable(conn, name, ...) ## S4 method for signature 'AzureKustoConnection,ANY' dbWriteTable(conn, name, value, method, ...) ## S4 method for signature 'AzureKustoConnection' dbCreateTable(conn, name, fields, ..., row.names = NULL, temporary = FALSE) ## S4 method for signature 'AzureKustoConnection,ANY' dbRemoveTable(conn, name, ...) ## S4 method for signature 'AzureKustoConnection' dbListTables(conn, ...) ## S4 method for signature 'AzureKustoConnection,ANY' dbExistsTable(conn, name, ...)
conn |
An AzureKustoConnection object. |
name |
A string containing a table name. |
... |
Further arguments passed to |
value |
For |
method |
For |
fields |
For |
row.names |
For |
temporary |
For |
These functions read, write, create and delete a table, list the tables in a Kusto database, and check for table existence. With the exception of dbWriteTable
, they ultimately call run_query
which does the actual work of communicating with the Kusto server. dbWriteTable
calls ingest_local
to write the data to the server; note that it only supports ingesting a local data frame, as per the DBI spec.
Kusto does not have the concept of temporary tables, so calling dbCreateTable
with temporary
set to anything other than FALSE
will generate an error.
dbReadTable
and dbWriteTable
are likely to be of limited use in practical scenarios, since Kusto tables tend to be much larger than available memory.
For dbReadTable
, an in-memory data frame containing the table.
AzureKusto-connection, dbConnect, run_query, ingest_local
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.location.kusto.windows.net", database="database"...) DBI::dbListTables(db) if(!DBI::dbExistsTable(db, "mtcars")) DBI::dbCreateTable(db, "mtcars") DBI::dbWriteTable(db, "mtcars", mtcars, method="inline") DBI::dbReadTable(db, "mtcars") DBI::dbRemoveTable(db, "mtcars") ## End(Not run)
## Not run: db <- DBI::dbConnect(AzureKusto(), server="https://mycluster.location.kusto.windows.net", database="database"...) DBI::dbListTables(db) if(!DBI::dbExistsTable(db, "mtcars")) DBI::dbCreateTable(db, "mtcars") DBI::dbWriteTable(db, "mtcars", mtcars, method="inline") DBI::dbReadTable(db, "mtcars") DBI::dbRemoveTable(db, "mtcars") ## End(Not run)
Method for the AzureRMR::az_resource_group class.
delete_kusto_cluster(name, confirm=TRUE, wait=FALSE)
name
: The name of the cluster.
confirm
: Whether to ask for confirmation before deleting.
wait
: Whether to wait until the deletion is complete.
NULL on successful deletion.
create_kusto_cluster, get_kusto_cluster, az_kusto
Kusto/Azure Data Explorer documentation
## Not run: rg <- AzureRMR::az_rm$ new(tenant="myaadtenant.onmicrosoft.com", app="app_id", password="password")$ get_subscription("subscription_id")$ get_resource_group("rgname") # delete a Kusto cluster rg$delete_kusto_cluster("mycluster") ## End(Not run)
## Not run: rg <- AzureRMR::az_rm$ new(tenant="myaadtenant.onmicrosoft.com", app="app_id", password="password")$ get_subscription("subscription_id")$ get_resource_group("rgname") # delete a Kusto cluster rg$delete_kusto_cluster("mycluster") ## End(Not run)
Escape/quote a string.
escape(x, parens = NA, collapse = " ") kql_vector(x, parens = NA, collapse = " ")
escape(x, parens = NA, collapse = " ") kql_vector(x, parens = NA, collapse = " ")
x |
An object to escape. Existing kql vectors will be left as is,
character vectors are escaped with single quotes, numeric vectors have
trailing |
parens , collapse
|
Controls behaviour when multiple values are supplied.
Default behaviour: lists are always wrapped in parens and separated by commas, identifiers are separated by commas and never wrapped, atomic vectors are separated by spaces and wrapped in parens if needed. |
Execute the Kusto query and export the result to Azure Storage.
Execute the Kusto query and export the result to Azure Storage.
export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... ) ## S3 method for class 'kusto_database_endpoint' export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... ) ## S3 method for class 'tbl_kusto' export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... )
export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... ) ## S3 method for class 'kusto_database_endpoint' export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... ) ## S3 method for class 'tbl_kusto' export( tbl, storage_uri, query = NULL, name_prefix = "export", key = "impersonate", format = "parquet", distributed = FALSE, ... )
tbl |
A Kusto database endpoint object, as returned by |
storage_uri |
The Azure Storage URI to export files to. |
query |
A Kusto query string |
name_prefix |
The filename prefix to use for exported files. |
key |
default "impersonate" which uses the account signed into Kusto to authenticate to Azure Storage. An Azure Storage account key. |
format |
Options are "parquet", "csv", "tsv", "json" |
distributed |
logical, indicates whether Kusto should distributed the export job to multiple nodes, in which case multiple files will be written to storage concurrently. |
... |
needed for agreement with generic. Not otherwise used. |
Walks the tree of ops and builds a stack.
flatten_query(op, ops = list())
flatten_query(op, ops = list())
op |
the current operation |
ops |
the stack of operations to append to, recursively |
Method for the AzureRMR::az_resource_group class.
get_kusto_cluster(name, location, node_size="D14_v2")
name
: The name of the cluster.
This method retrieves an existing Kusto cluster resource.
An object of class az_kusto
representing the created cluster.
create_kusto_cluster, delete_kusto_cluster, az_kusto
Kusto/Azure Data Explorer documentation
## Not run: rg <- AzureRMR::get_azure_login("myaadtenant")$ get_subscription("subscription_id")$ get_resource_group("rgname") # get a Kusto cluster rg$get_kusto_cluster("mykusto") ## End(Not run)
## Not run: rg <- AzureRMR::get_azure_login("myaadtenant")$ get_subscription("subscription_id")$ get_resource_group("rgname") # get a Kusto cluster rg$get_kusto_cluster("mykusto") ## End(Not run)
Manage AAD authentication tokens for Kusto clusters
get_kusto_token( server = NULL, clustername, location = NULL, tenant = NULL, app = .kusto_app_id, auth_type = NULL, version = 2, ... ) delete_kusto_token( server = NULL, clustername, location = NULL, tenant = NULL, app = .kusto_app_id, auth_type = NULL, version = 2, ..., hash = NULL, confirm = TRUE ) list_kusto_tokens()
get_kusto_token( server = NULL, clustername, location = NULL, tenant = NULL, app = .kusto_app_id, auth_type = NULL, version = 2, ... ) delete_kusto_token( server = NULL, clustername, location = NULL, tenant = NULL, app = .kusto_app_id, auth_type = NULL, version = 2, ..., hash = NULL, confirm = TRUE ) list_kusto_tokens()
server |
The URI of your Kusto cluster. If not supplied, it is obtained from the |
clustername |
The cluster name. |
location |
The cluster location. Leave this blank for a Microsoft-internal Kusto cluster like "help". |
tenant |
Your Azure Active Directory (AAD) tenant. Can be a GUID, a name ("myaadtenant") or a fully qualified domain name ("myaadtenant.com"). |
app |
The ID of the Azure Active Directory app/service principal to authenticate with. Defaults to the ID of the KustoClient app. |
auth_type |
The authentication method to use. Can be one of "authorization_code", "device_code", "client_credentials" or "resource_owner". The default is to pick one based on the other arguments. |
version |
The AAD version to use. There should be no reason to change this from the default value of 2. |
... |
Other arguments to pass to AzureAuth::get_azure_token. |
hash |
For |
confirm |
For |
get_kusto_token
returns an authentication token for the given cluster, caching its value on disk. delete_kusto_token
deletes a cached token, and list_kusto_tokens
lists all cached tokens.
By default, authentication tokens will be obtained using the main KustoClient Active Directory app. This app can be used to authenticate with any Kusto cluster (assuming, of course, you have the proper credentials).
get_kusto_token
returns an object of class AzureAuth::AzureToken representing the authentication token, while list_kusto_tokens
returns a list of such objects. delete_azure_token
returns NULL on a successful delete.
kusto_database_endpoint, AzureAuth::get_azure_token
## Not run: get_kusto_token("https://myclust.australiaeast.kusto.windows.net") get_kusto_token(clustername="myclust", location="australiaeast") # authenticate using client_credentials method: see ?AzureAuth::get_azure_token get_kusto_token("https://myclust.australiaeast.kusto.windows.net", tenant="mytenant", app="myapp", password="password") ## End(Not run)
## Not run: get_kusto_token("https://myclust.australiaeast.kusto.windows.net") get_kusto_token(clustername="myclust", location="australiaeast") # authenticate using client_credentials method: see ?AzureAuth::get_azure_token get_kusto_token("https://myclust.australiaeast.kusto.windows.net", tenant="mytenant", app="myapp", password="password") ## End(Not run)
Flag a character string as a Kusto identifier
ident(...)
ident(...)
... |
character strings to flag as Kusto identifiers |
Pass an already-escaped string to Kusto
ident_q(...)
ident_q(...)
... |
character strings to treat as already-escaped identifiers |
Ingestion functions for Kusto
ingest_local( database, src, dest_table, method = NULL, staging_container = NULL, ingestion_token = database$token, http_status_handler = "stop", ... ) ingest_url(database, src, dest_table, async = FALSE, ...) ingest_blob( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... ) ingest_adls2( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... ) ingest_adls1( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... )
ingest_local( database, src, dest_table, method = NULL, staging_container = NULL, ingestion_token = database$token, http_status_handler = "stop", ... ) ingest_url(database, src, dest_table, async = FALSE, ...) ingest_blob( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... ) ingest_adls2( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... ) ingest_adls1( database, src, dest_table, async = FALSE, key = NULL, token = NULL, sas = NULL, ... )
database |
A Kusto database endpoint object, created with kusto_database_endpoint. |
src |
The source data. This can be either a data frame, local filename, or URL. |
dest_table |
The name of the destination table. |
method |
For local ingestion, the method to use. See 'Details' below. |
staging_container |
For local ingestion, an Azure storage container to use for staging the dataset. This can be an object of class either AzureStor::blob_container or AzureStor::adls_filesystem. Only used if |
ingestion_token |
For local ingestion, an Azure Active Directory authentication token for the cluster ingestion endpoint. Only used if |
http_status_handler |
For local ingestion, how to handle HTTP conditions >= 300. Defaults to "stop"; alternatives are "warn", "message" and "pass". The last option will pass through the raw response object from the server unchanged, regardless of the status code. This is mostly useful for debugging purposes, or if you want to see what the Kusto REST API does. Only used if |
... |
Named arguments to be treated as ingestion parameters. |
async |
For the URL ingestion functions, whether to do the ingestion asychronously. If TRUE, the function will return immediately while the server handles the operation in the background. |
key , token , sas
|
Authentication arguments for the Azure storage ingestion methods. If multiple arguments are supplied, a key takes priority over a token, which takes priority over a SAS. Note that these arguments are for authenticating with the Azure storage account, as opposed to Kusto itself. |
There are up to 3 possible ways to ingest a local dataset, specified by the method
argument.
method="indirect"
: The data is uploaded to blob storage, and then ingested from there. This is the default if the AzureStor package is present.
method="streaming"
: The data is uploaded to the cluster ingestion endpoint. This is the default if the AzureStor package is not present, however be aware that currently (as of February 2019) streaming ingestion is in beta and has to be enabled for a cluster by filing a support ticket.
method="inline"
: The data is embedded into the command text itself. This is only recommended for testing purposes, or small datasets.
Note that the destination table must be created ahead of time for the ingestion to proceed.
## Not run: # ingesting from local: # ingest via Azure storage cont <- AzureStor::storage_container("https://mystorage.blob.core.windows.net/container", sas="mysas") ingest_local(db, "file.csv", "table", method="indirect", storage_container=cont) ingest_local(db, "file.csv", "table", method="streaming") # ingest by inlining data into query ingest_inline(db, "file.csv", "table", method="inline") # ingesting online data: # a public dataset: Microsoft web data from UCI machine learning repository ingest_url(db, "https://archive.ics.uci.edu/ml/machine-learning-databases/anonymous/anonymous-msweb.data", "table") # from blob storage: ingest_blob(db, "https://mystorage.blob.core.windows.net/container/myblob", "table", sas="mysas") # from ADLSGen2: token <- AzureRMR::get_azure_token("https://storage.azure.com", "mytenant", "myapp", "password") ingest_blob(db, "abfss://[email protected]/data/myfile", "table", token=token) ## End(Not run)
## Not run: # ingesting from local: # ingest via Azure storage cont <- AzureStor::storage_container("https://mystorage.blob.core.windows.net/container", sas="mysas") ingest_local(db, "file.csv", "table", method="indirect", storage_container=cont) ingest_local(db, "file.csv", "table", method="streaming") # ingest by inlining data into query ingest_inline(db, "file.csv", "table", method="inline") # ingesting online data: # a public dataset: Microsoft web data from UCI machine learning repository ingest_url(db, "https://archive.ics.uci.edu/ml/machine-learning-databases/anonymous/anonymous-msweb.data", "table") # from blob storage: ingest_blob(db, "https://mystorage.blob.core.windows.net/container/myblob", "table", sas="mysas") # from ADLSGen2: token <- AzureRMR::get_azure_token("https://storage.azure.com", "mytenant", "myapp", "password") ingest_blob(db, "abfss://[email protected]/data/myfile", "table", token=token) ## End(Not run)
These methods are the same as other joining methods, with the exception of the .strategy
, .shufflekeys
and .num_partitions
optional arguments. They provide hints to the Kusto engine on how to execute the join, and can sometimes be useful to speed up a query. See the Kusto documentation for more details.
## S3 method for class 'tbl_kusto_abstract' inner_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' left_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' right_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' full_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' semi_join( x, y, by = NULL, copy = NULL, ..., suffix = c(".x", ".y"), .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' anti_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL, ... )
## S3 method for class 'tbl_kusto_abstract' inner_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' left_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' right_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' full_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), ..., keep = NULL, .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' semi_join( x, y, by = NULL, copy = NULL, ..., suffix = c(".x", ".y"), .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL ) ## S3 method for class 'tbl_kusto_abstract' anti_join( x, y, by = NULL, copy = NULL, suffix = c(".x", ".y"), .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL, .remote = NULL, ... )
x , y
|
Kusto tbls. |
by |
The columns to join on. |
copy |
Needed for agreement with generic. Not otherwise used. |
suffix |
The suffixes to use for deduplicating column names. |
... |
Other arguments passed to lower-level functions. |
keep |
Needed for agreement with generic. Not otherwise used. Kusto retains keys from both sides of joins. |
.strategy |
A join strategy hint to pass to Kusto. Currently the values supported are "shuffle" and "broadcast". |
.shufflekeys |
A character vector of column names to use as shuffle keys. |
.num_partitions |
The number of partitions for a shuffle query. |
.remote |
A join strategy hint to use for cross-cluster joins. Can be "left", "right", "local" or "auto" (the default). |
## Not run: tbl1 <- tbl_kusto(db, "table1") tbl2 <- tbl_kusto(db, "table2") # standard dplyr syntax: left_join(tbl1, tbl2) # Kusto extensions: left_join(tbl1, tbl2, .strategy = "broadcast") # a broadcast join left_join(tbl1, tbl2, .shufflekeys = c("var1", "var2")) # shuffle join with shuffle keys left_join(tbl1, tbl2, .num_partitions = 5) # no. of partitions for a shuffle join ## End(Not run)
## Not run: tbl1 <- tbl_kusto(db, "table1") tbl2 <- tbl_kusto(db, "table2") # standard dplyr syntax: left_join(tbl1, tbl2) # Kusto extensions: left_join(tbl1, tbl2, .strategy = "broadcast") # a broadcast join left_join(tbl1, tbl2, .shufflekeys = c("var1", "var2")) # shuffle join with shuffle keys left_join(tbl1, tbl2, .num_partitions = 5) # no. of partitions for a shuffle join ## End(Not run)
These functions test whether an object is of the given class.
is_kusto_database(x) is_kusto_cluster(x)
is_kusto_database(x) is_kusto_cluster(x)
x |
An R object. |
Tag character strings as Kusto Query Language. Assumes the string is valid and properly escaped.
kql(...)
kql(...)
... |
character strings to tag as KQL |
Return a function representing a KQL aggregation function
kql_aggregate(f)
kql_aggregate(f)
f |
Name of the Kusto aggregation function |
Build the tbl object into a data structure representing a Kusto query
kql_build(op)
kql_build(op)
op |
A nested sequence of query operations, i.e. tbl_kusto$ops |
dplyr's mutate verb can include aggregations, but Kusto's extend verb cannot. If the mutate contains no aggregations, then it can emit an extend clause. If the mutate contains an aggregation and the tbl is ungrouped, then it must emit a summarize clause grouped by all variables. If the mutate contains an aggregation and the tbl is grouped, then it must join to a subquery containing the summarize clause.
## S3 method for class 'op_mutate' kql_build(op, ...)
## S3 method for class 'op_mutate' kql_build(op, ...)
op |
A nested sequence of query operations, i.e. tbl_kusto$ops |
... |
Needed for agreement with generic. Not otherwise used. |
Escape a Kusto identifier with [' ']
kql_escape_ident(x)
kql_escape_ident(x)
x |
An identifier to escape |
Pass through an already-escaped Kusto identifier
kql_escape_ident_q(x)
kql_escape_ident_q(x)
x |
An identifier to pass through |
Escape a Kusto logical value. Converts TRUE/FALSE to true / false
kql_escape_logical(x)
kql_escape_logical(x)
x |
A logical value to escape |
Escape a Kusto string by single-quoting
kql_escape_string(x)
kql_escape_string(x)
x |
A string to escape |
Return a function representing a scalar KQL infix operator
kql_infix(f)
kql_infix(f)
f |
Name of a Kusto infix operator / function |
Return a function representing a scalar KQL prefix function
kql_prefix(f, n = NULL)
kql_prefix(f, n = NULL)
f |
Name of a Kusto infix function |
n |
Number of arguments accepted by the Kusto prefix function |
Render a set of operations on a tbl_kusto_abstract to a Kusto query
kql_render(query, ...)
kql_render(query, ...)
query |
The tbl_kusto instance with a sequence of operations in $ops |
... |
needed for agreement with generic. Not otherwise used. |
Build a kql_variant class out of the environments holding scalar and aggregation function definitions
kql_translate_env()
kql_translate_env()
Builds an environment from a list of R -> Kusto query language translation pairs.
kql_translator(..., .funs = list(), .parent = new.env(parent = emptyenv()))
kql_translator(..., .funs = list(), .parent = new.env(parent = emptyenv()))
... |
Pairs of R call = Kusto call translations as individual arguments |
.funs |
Parse of R call = Kusto call translations in list format |
.parent |
A parent environment to attach this env onto |
Return a function representing a KQL window function
kql_window(f)
kql_window(f)
f |
Name of the Kusto aggregation function |
Endpoints for communicating with a Kusto database
kusto_database_endpoint( ..., .connection_string = NULL, .query_token = NULL, .use_integer64 = FALSE )
kusto_database_endpoint( ..., .connection_string = NULL, .query_token = NULL, .use_integer64 = FALSE )
... |
Named arguments which are the properties for the endpoint object. See 'Details' below for the properties that AzureKusto recognises. |
.connection_string |
An alternative way of specifying the properties, as a database connection string. Properties supplied here override those in |
.query_token |
Optionally, an Azure Active Directory (AAD) token to authenticate with. If this is supplied, it overrides other tokens specified in |
.use_integer64 |
For |
This is a list of properties recognised by kusto_database_endpoint
, and their alternate names. Property names not in this list will generate an error. Note that not all properties that are recognised are currently supported by AzureKusto.
General properties:
server: The URI of the server, usually of the form 'https://clustername.location.kusto.windows.net'.
addr, address, network address, datasource, host
database: The database.
initialcatalog, dbname
tenantid: The AAD tenant name or ID to authenticate with.
authority
appclientid: The AAD app/service principal ID
applicationclientid
traceclientversion: The client version for tracing.
queryconsistency: The level of query consistency. Defaults to "weakconsistency".
response_dynamic_serialization: How to serialize dynamic responses.
response_dynamic_serialization_2: How to serialize dynamic responses.
User authentication properties:
password
user: The user name.
uid, userid
traceusername: The user name for tracing.
usertoken: The AAD token for user authentication.
usertoken, usrtoken
fed: Logical, whether federated authentication is enabled. Currently unsupported; if this is TRUE, kusto_database_endpoint
will print a warning and ignore it.
federated security, federated, aadfed, aadfederatedsecurity
App authentication properties:
appkey: The secret key for the app.
applicationkey
traceappname: The AAD app for tracing.
apptoken: The AAD token for app authentication.
apptoken, applicationtoken
Currently, AzureKusto only supports authentication via Azure Active Directory. Authenticating with DSTS is planned for the future.
The way kusto_database_endpoint
obtains an AAD token is as follows.
If the .query_token
argument is supplied, use it.
Otherwise, if the usertoken
property is supplied, use it.
Otherwise, if the apptoken
property is supplied, use it.
Otherwise, if the appclientid
property is supplied, use it to obtain a token:
With the user
and pwd
properties if available
Or with the appkey
property if available
Otherwise do an interactive authentication and ask for the user credentials
Otherwise, if no appclientid
property is supplied, authenticate with the KustoClient app:
With the user
and pwd
properties if available
Otherwise do an interactive authentication and ask for the user credentials using a device code
An object of class kusto_database_endpoint
.
## Not run: kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1") # supplying a token obtained previously token <- get_kusto_token("myclust.australiaeast.kusto.windows.net") kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1", .query_token=token) ## End(Not run)
## Not run: kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1") # supplying a token obtained previously token <- get_kusto_token("myclust.australiaeast.kusto.windows.net") kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1", .query_token=token) ## End(Not run)
Execute the query, store the results in a table, and return a reference to the new table Run a Kusto query and export results to Azure Storage in Parquet or CSV format.
kusto_export_cmd(query, storage_uri, name_prefix, key, format, distributed)
kusto_export_cmd(query, storage_uri, name_prefix, key, format, distributed)
query |
The text of the Kusto query to run |
storage_uri |
The URI of the blob storage container to export to |
name_prefix |
The filename prefix for each exported file |
key |
The account key for the storage container. uses the identity that is signed into Kusto to authenticate to Azure Storage. |
format |
Options are "parquet", "csv", "tsv", "json" |
distributed |
logical, indicates whether Kusto should distributed the export job to multiple nodes, in which case multiple files will be written to storage concurrently. |
AzureKusto implements a subset of the DBI specification for interfacing with databases in R. The following methods are supported:
Connections: dbConnect, dbDisconnect, dbCanConnect
Table management: dbExistsTable, dbCreateTable, dbRemoveTable, dbReadTable, dbWriteTable
Querying: dbGetQuery, dbSendQuery, dbFetch, dbSendStatement, dbExecute, dbListFields, dbColumnInfo
Kusto is quite different to the SQL databases that DBI targets, which affects the behaviour of certain DBI methods and renders other moot.
Kusto is connectionless. dbConnect
simply wraps a database endpoint object, created with kusto_database_endpoint. Similarly, dbDisconnect
always returns TRUE. dbCanConnect
attempts to check if querying the database will succeed, but this may not be accurate.
Temporary tables are not a Kusto concept, so dbCreateTable(*, temporary=TRUE)
will throw an error.
It only supports synchronous queries, with a default timeout of 4 minutes. dbSendQuery
and dbSendStatement
will wait for the query to execute, rather than returning immediately. The object returned contains the full result of the query, which dbFetch
extracts.
The Kusto Query Language (KQL) is not SQL, and so higher-level SQL methods are not implemented.
This method collapses a column into a list
## S3 method for class 'tbl_kusto_abstract' nest(.data, ...)
## S3 method for class 'tbl_kusto_abstract' nest(.data, ...)
.data |
A kusto tbl. |
... |
Specification of columns to nest. Translates to summarize make_list() in Kusto. |
The "base case" operation representing the tbl itself and its column variables
op_base(x, vars, class = character())
op_base(x, vars, class = character())
x |
A tbl object |
vars |
A vector of column variables in the tbl |
class |
The class that op_base should inherit from, default is character() |
A double-table verb, e.g. joins, setops
op_double(name, x, y, args = list())
op_double(name, x, y, args = list())
name |
The name of the operation, e.g. 'left_join', 'union_all' |
x |
The "left" tbl |
y |
The "right" tbl |
args |
Other arguments passed to the operation verb |
Look up the applicable grouping variables for an operation based on the data source and preceding sequence of operations
op_grps(op)
op_grps(op)
op |
An operation instance |
A class representing a single-table verb
op_single(name, x, dots = list(), args = list())
op_single(name, x, dots = list(), args = list())
name |
the name of the operation verb, e.g. "select", "filter" |
x |
the tbl object |
dots |
expressions passed to the operation verb function |
args |
other arguments passed to the operation verb function |
Look up the applicable variables in scope for a given operation based on the data source and preceding sequence of operations
op_vars(op)
op_vars(op)
op |
An operation instance |
Run a query or command against a Kusto database
run_query(database, qry_cmd, ..., .http_status_handler = "stop")
run_query(database, qry_cmd, ..., .http_status_handler = "stop")
database |
A Kusto database endpoint object, as returned by |
qry_cmd |
A string containing the query or command. In KQL, a database management command is a statement that starts with a "." |
... |
Named arguments to be used as parameters for a parameterized query. These are ignored for database management commands. |
.http_status_handler |
The function to use to handle HTTP status codes. The default "stop" will throw an R error via |
This function is the workhorse of the AzureKusto package. It communicates with the Kusto server and returns the query or command results, as data frames.
kusto_database_endpoint, ingest_local, ingest_url, ingest_blob, ingest_adls2
## Not run: endp <- kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1") # a command run_query(endp, ".show table iris") # a query run_query(endp, "iris | count") ## End(Not run)
## Not run: endp <- kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1") # a command run_query(endp, ".show table iris") # a query run_query(endp, "iris | count") ## End(Not run)
Translate a sequence of dplyr operations on a tbl into a Kusto query string.
## S3 method for class 'tbl_kusto_abstract' show_query(x, ...)
## S3 method for class 'tbl_kusto_abstract' show_query(x, ...)
x |
A tbl_kusto or tbl_kusto_abstract instance |
... |
needed for agreement with generic. Not otherwise used. |
This method is the same as other summarise methods, with the exception of the .strategy
, .shufflekeys
and .num_partitions
optional arguments. They provide hints to the Kusto engine on how to execute the summarisation, and can sometimes be useful to speed up a query. See the Kusto documentation for more details.
## S3 method for class 'tbl_kusto_abstract' summarise( .data, ..., .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL )
## S3 method for class 'tbl_kusto_abstract' summarise( .data, ..., .strategy = NULL, .shufflekeys = NULL, .num_partitions = NULL )
.data |
A Kusto tbl. |
... |
Summarise expressions. |
.strategy |
A summarise strategy to pass to Kusto. Currently the only value supported is "shuffle". |
.shufflekeys |
A character vector of column names to use as shuffle keys. |
.num_partitions |
The number of partitions for a shuffle query. |
## Not run: tbl1 <- tbl_kusto(db, "table1") ## standard dplyr syntax: summarise(tbl1, mx = mean(x)) ## Kusto extensions: summarise(tbl1, mx = mean(x), .strategy = "broadcast") # a broadcast summarise summarise(tbl1, mx = mean(x), .shufflekeys = c("var1", "var2")) # a shuffle summarise with keys summarise(tbl1, mx = mean(x), .num_partitions = 5) # no. of partitions for a shuffle summarise ## End(Not run)
## Not run: tbl1 <- tbl_kusto(db, "table1") ## standard dplyr syntax: summarise(tbl1, mx = mean(x)) ## Kusto extensions: summarise(tbl1, mx = mean(x), .strategy = "broadcast") # a broadcast summarise summarise(tbl1, mx = mean(x), .shufflekeys = c("var1", "var2")) # a shuffle summarise with keys summarise(tbl1, mx = mean(x), .num_partitions = 5) # no. of partitions for a shuffle summarise ## End(Not run)
A tbl object representing a table in a Kusto database.
tbl_kusto(kusto_database, table_name, ...)
tbl_kusto(kusto_database, table_name, ...)
kusto_database |
An instance of kusto_database_endpoint that this table should be queried from |
table_name |
The name of the table in the Kusto database |
... |
parameters to pass in case the Kusto source table is a parameterized function. |
Translate R expressions into Kusto Query Language equivalents.
translate_kql(...)
translate_kql(...)
... |
Expressions to translate. |
This method takes a list column and expands it so that each element of the list gets its own row. unnest() translates to Kusto's mv-expand operator.
## S3 method for class 'tbl_kusto_abstract' unnest( data, cols, ..., keep_empty = FALSE, ptype = NULL, names_sep = NULL, names_repair = NULL, .drop = NULL, .id = NULL, .sep = NULL, .preserve = NULL )
## S3 method for class 'tbl_kusto_abstract' unnest( data, cols, ..., keep_empty = FALSE, ptype = NULL, names_sep = NULL, names_repair = NULL, .drop = NULL, .id = NULL, .sep = NULL, .preserve = NULL )