| | |
- ReadNOS(data=None, location=None, access_id=None, access_key=None, buffer_size=16, return_type='NOSREAD_RECORD', sample_perc=1.0, stored_as='TEXTFILE', full_scan=False, manifest=False, row_format=None, header=True, **generic_arguments)
- DESCRIPTION:
ReadNOS() function enables access to external files in JSON, CSV, or Parquet format.
You must have the EXECUTE FUNCTION privilege on TD_SYSFNLIB.READ_NOS.
PARAMETERS:
data:
Optional Argument.
Specifies the input teradataml DataFrame.
Types: teradataml DataFrame
location:
Optional Argument.
Specifies the location value, which is a Uniform Resource Identifier
(URI) pointing to the data in the external object storage system. The
location value includes the following components:
Amazon S3:
/connector/bucket.endpoint/[key_prefix].
Azure Blob storage and Azure Data Lake Storage Gen2:
/connector/container.endpoint/[key_prefix].
Google Cloud Storage:
/connector/endpoint/bucket/[key_prefix].
connector:
Identifies the type of external storage system where the data is located.
Teradata requires the storage location to start with the following for
all external storage locations:
* Amazon S3 storage location must begin with /S3 or /s3
* Azure Blob storage location (including Azure Data Lake Storage Gen2
in Blob Interop Mode) must begin with /AZ or /az.
* Google Cloud Storage location must begin with /GS or /gs.
storage-account:
Used by Azure. The Azure storage account contains your
Azure storage data objects.
endpoint:
A URL that identifies the system-specific entry point for
the external object storage system.
bucket (Amazon S3, Google Cloud Storage) or container (Azure Blob
storage and Azure Data Lake Storage Gen2):
A container that logically groups stored objects in the
external storage system.
key_prefix:
Identifies one or more objects in the logical organization of
the bucket data. Because it is a key prefix, not an actual
directory path, the key prefix may match one or more objects
in the external storage. For example, the key prefix
"/fabrics/cotton/colors/b/" would match objects
/fabrics/cotton/colors/blue, /fabrics/cotton/colors/brown, and
/fabrics/cotton/colors/black. If there are organization levels below
those, such as /fabrics/cotton/colors/blue/shirts, the same key
prefix would gather those objects too.
Note:
Vantage validates only the first file it encounters from the
location key prefix.
For example, this location value might specify all objects on an
Amazon cloud storage system for the month of December, 2001:
location = "/S3/YOUR-BUCKET.s3.amazonaws.com/csv/US-Crimes/csv-files/2001/Dec/"
connector: S3
bucket: YOUR-BUCKET
endpoint: s3.amazonaws.com
key_prefix: csv/US-Crimes/csv-files/2001/Dec/
Following location could specify an individual storage object (or file), Day1.csv:
location = "/S3/YOUR-BUCKET.s3.amazonaws.com/csv/US-Crimes/csv-files/2001/Dec/Day1.csv"
connector: S3
bucket: YOUR-BUCKET
endpoint: s3.amazonaws.com
key_prefix: csv/US-Crimes/csv-files/2001/Dec/Day1.csv
Following location specifies an entire container in an Azure external
object store (Azure Blob storage or Azure Data Lake Storage Gen2).
The container may contain multiple file objects:
location = "/AZ/YOUR-STORAGE-ACCOUNT.blob.core.windows.net/nos-csv-data"
connector: AZ
bucket: YOUR-STORAGE-ACCOUNT
endpoint: blob.core.windows.net
key_prefix: nos-csv-data
This is an example of a Google Cloud Storage location:
location = "/gs/storage.googleapis.com/YOUR-BUCKET/CSVDATA/RIVERS/rivers.csv"
connector: GS
bucket: YOUR-BUCKET
endpoint: storage.googleapis.com,
key_prefix: CSVDATA/RIVERS/rivers.csv
Types: str
access_id:
Optional Argument.
Specifies the identification for accessing external storage.
This option is not necessary if an authentication object is defined,
for example, in the EXTERNAL SECURITY clause of a function mapping
or if IAM roles are defined for S3 repositories.
See "CREATE FUNCTION MAPPING" in Teradata Vantage™ - SQL Data
Definition Language Syntax and Examples, B035-1144.
To set the function mapping for ReadNOS() function, set the following
configuration options with function mapping name.
teradataml.options.configure.read_nos_function_mapping = "<function mapping name>"
Types: str
access_key:
Optional Argument.
Specifies the password for accessing external storage.
This option is not necessary if an authentication object is defined,
for example, in the EXTERNAL SECURITY clause of a function mapping
or if IAM roles are defined for S3 repositories.
See "CREATE FUNCTION MAPPING" in Teradata Vantage - SQL Data
Definition Language Syntax and Examples.
To set the function mapping for ReadNOS() function, set the following
configuration options with function mapping name.
teradataml.options.configure.read_nos_function_mapping = "<function mapping name>"
Types: str
buffer_size:
Optional Argument.
Specifies the size of the network buffer to allocate when retrieving
data from the external storage repository. 16 MB, which is the maximum
value.
Default Value: 16
Types: int
return_type:
Optional Argument.
Specifies the format in which data is returned.
Permitted Values:
* NOSREAD_RECORD:
Returns one row for each external record along with its metadata.
Access external records by specifying one of the following:
- Input teradataml DataFrame, location, and teradataml DataFrame
on an empty table. For CSV, you can include a schema definition.
- Input teradataml DataFrame with a row for each external file. For
CSV, this method does not support a schema definition.
For an empty single-column input table, do the following:
- Define an input teradataml DataFrame with a single column, Payload,
with the appropriate data type:
JSON and CSV
This column determines the output Payload column return type.
- Specify the filepath in the "location" argument.
For a multiple-column input table, define an input teradataml
DataFrame with the following columns:
+------------------+-------------------------------------+
| Column Name | Data Types |
+------------------+-------------------------------------+
| Location | VARCHAR(2048) CHARACTER SET UNICODE |
+------------------+-------------------------------------+
| ObjectVersionID | VARCHAR(1024) CHARACTER SET UNICODE |
+------------------+-------------------------------------+
| OffsetIntoObject | BIGINT |
+------------------+-------------------------------------+
| ObjectLength | BIGINT |
+------------------+-------------------------------------+
| Payload | JSON |
+------------------+-------------------------------------+
| CSV | VARCHAR |
+------------------+-------------------------------------+
This teradataml DataFrame can be populated using the output of the
'NOSREAD_KEYS' return type.
* NOSREAD_KEYS:
Retrieve the list of files from the path specified in the "location" argument.
A schema definition is not necessary.
'NOSREAD_KEYS' returns Location, ObjectVersionID, ObjectTimeStamp,
and ObjectLength (size of external file).
* NOSREAD_PARQUET_SCHEMA:
Returns information about the Parquet data schema. If you are using
the "full_scan" option, use 'NOSREAD_PARQUET_SCHEMA'; otherwise you can use
'NOSREAD_SCHEMA' to get information about the Parquet schema. For information
about the mapping between Parquet data types and Teradata data types, see
Parquet External Files in Teradata Vantage - SQL Data Definition Language
Syntax and Examples.
* NOSREAD_RAW:
Retrieves file data from the external storage services, not specific records.
Retrieved data is returned as CLOB/BLOB. You can retrieve a complete file from
external storage and save in Teradata CLOB/BLOB format. The maximum amount of
data that can be retrieved from the external storage and saved in the DataFrame
column is 2GB, the Vantage limit for LOBs. The ObjectLength corresponds to the
length of CLOB/BLOB column read from the external storage.
This information is provided in the form of a DataFrame returned to the ReadNOS()
function. The Payload column in the input data is only used to determine the
datatype of the column in which the returned data is stored.
Define the input data with the following columns:
- Location VARCHAR(2048) CHARACTER SET UNICODE
- ObjectVersionID VARCHAR(1024) CHARACTER SET UNICODE
- ObjectTimeStamp TIMESTAMP(6)
- OffsetIntoObject BIGINT
- BytesToRead BIGINT
- Payload CLOB/BLOB
ReadNOS() function returns a DataFrame with the following columns:
- Location VARCHAR(2048) CHARACTER SET UNICODE
- ObjectVersionID VARCHAR(1024) CHARACTER SET UNICODE
- ObjectTimeStamp TIMESTAMP(6)
- OffsetIntoObject BIGINT
- ObjectLength BIGINT
- Payload CLOB/BLOB, based on input table CLOB/BLOB Column.
Default Value: "NOSREAD_RECORD"
Types: str
sample_perc:
Optional Argument.
Specifies the percentage of rows to retrieve from the external
storage repository when "return_type" is 'NOSREAD_RECORD'. The valid
range of values is from 0.0 to 1.0, where 1.0 represents 100%
of the rows.
Default Value: 1.0
Types: float
stored_as:
Optional Argument.
Specifies the formatting style of the external data.
Permitted Values:
* PARQUET-
The external data is formatted as Parquet. This is a
required parameter for Parquet data.
* TEXTFILE-
The external data uses a text-based format, such as
CSV or JSON.
Default Value: "TEXTFILE"
Types: str
full_scan:
Optional Argument.
Specifies whether ReadNOS() function scans columns of variable length
types (CHAR, VARCHAR, BYTE, VARBYTE, JSON, and BSON) to discover the
maximum length.
When set to True, the size of variable length data is determined
from the Parquet data.
Note:
Choosing this value can impact performance because all variable length
data type columns in each Parquet file at the location must be scanned
to assess the value having the greatest length.
When set to False, variable length field sizes are assigned the
Vantage maximum value for the particular data type.
Note:
"full_scan" is only used with a "return_type" of 'NOSREAD_PARQUET_SCHEMA'.
Default Value: False
Types: bool
manifest:
Optional Argument.
Specifies whether the location value points to a manifest file (a
file containing a list of files to read) or object name. The object
name can include the full path or a partial path. It must identify a
single file containing the manifest.
Note:
Individual entries within the manifest file must show
complete paths.
Below is an example of a manifest file that contains a list of locations
in JSON format.
{
"entries": [
{"url":"s3://nos-core-us-east-1/UNICODE/JSON/mln-key/data-10/data-8_9_02-10.json"},
{"url":"s3://nos-core-us-east-1/UNICODE/JSON/mln-key/data-10/data-8_9_02-101.json"},
{"url":"s3://nos-core-us-east-1/UNICODE/JSON/mln-key/data-10/data-10-01/data-8_9_02-102.json"},
{"url":"s3://nos-core-us-east-1/UNICODE/JSON/mln-key/data-10/data-10-01/data-8_9_02-103.json"}
]
}
Default Value: False
Types: bool
row_format:
Optional Argument.
Specifies the encoding format of the external row.
For example:
row_format = {'field_delimiter': ',', 'record_delimiter': '\n', 'character_set': 'LATIN'}
If string value is used, JSON format must be used to specify the row format.
For example:
row_format = '{"field_delimiter": ",", "record_delimiter": "\n", "character_set": "LATIN"}'
Format can include only the three keys shown above. Key names and values
are case-specific, except for the value for "character_set", which can
use any combination of letter cases.
The character set specified in "row_format" must be compatible with the character
set of the Payload column.
Do not specify "row_format" for Parquet format data.
For a JSON column, these are the default values:
UNICODE:
row_format = {"record_delimiter":"\n", "character_set":"UTF8"}
LATIN:
row_format = {"record_delimiter":"\n", "character_set":"LATIN"}
For a CSV column, these are the default values:
UNICODE:
row_format = '{"character_set":"UTF8"}'
LATIN:
row_format = '{"character_set":"LATIN"}'
User can specify the following options:
field_delimiter: The default is "," (comma). User can also specify a
custom field delimiter, such as tab "\t".
record_delimiter: New line feed character: "\n". A line feed
is the only acceptable record delimiter.
character_set: "UTF8" or "LATIN". If you do not specify a "row_format"
or payload column, Vantage assumes UTF8 Unicode.
Types: str or dict
header:
Optional Argument.
Specifies whether the first row of data in an input CSV file is
interpreted as column headings for the subsequent rows of data. Use
this parameter only when a CSV input file is not associated with a
separate schema object that defines columns for the CSV data.
Default Value: True
Types: bool
**generic_arguments:
Specifies the generic keyword arguments SQLE functions accept.
Below are the generic keyword arguments:
persist:
Optional Argument.
Specifies whether to persist the results of the function in table or not.
When set to True, results are persisted in table; otherwise, results
are garbage collected at the end of the session.
Default Value: False
Types: boolean
volatile:
Optional Argument.
Specifies whether to put the results of the function in volatile table or not.
When set to True, results are stored in volatile table, otherwise not.
Default Value: False
Types: boolean
Function allows the user to partition, hash, order or local order the input
data. These generic arguments are available for each argument that accepts
teradataml DataFrame as input and can be accessed as:
* "<input_data_arg_name>_partition_column" accepts str or list of str (Strings)
* "<input_data_arg_name>_hash_column" accepts str or list of str (Strings)
* "<input_data_arg_name>_order_column" accepts str or list of str (Strings)
* "local_order_<input_data_arg_name>" accepts boolean
Note:
These generic arguments are supported by teradataml if the underlying Analytic Database
function supports, else an exception is raised.
RETURNS:
Instance of ReadNOS.
Output teradataml DataFrames can be accessed using attribute
references, such as ReadNOS.<attribute_name>.
Output teradataml DataFrame attribute name is:
result
RAISES:
TeradataMlException, TypeError, ValueError
EXAMPLES:
# Notes:
# 1. Get the connection to Vantage to execute the function.
# 2. One must import the required functions mentioned in
# the example from teradataml.
# 3. Function will raise error if not supported on the Vantage
# user is connected to.
# Check the list of available analytic functions.
display_analytic_functions()
# Example 1: Reading PARQUET file from AWS S3 location.
obj = ReadNOS(location="/S3/s3.amazonaws.com/Your-Bucket/location/",
access_id="YOUR-ID",
access_key="YOUR-KEY",
stored_as="PARQUET")
# print the result DataFame.
print(obj.result)
# Example 2: Read PARQUET file in external storage with one row for each external
# record along with its metadata.
obj = ReadNOS(location="/S3/s3.amazonaws.com/Your-Bucket/location/",
access_id="YOUR-ID",
access_key="YOUR-KEY",
return_type="NOSREAD_KEYS")
# print the result DataFame.
print(obj.result)
# Example 3: Read CSV file from external storage.
obj = ReadNOS(location="/S3/s3.amazonaws.com/Your-Bucket/csv-location/",
access_id="YOUR-ID",
access_key="YOUR-KEY",
stored_as="TEXTFILE")
# print the result DataFame.
print(obj.result)
# Example 4: Read CSV file in external storage using "data" argument.
# Create a table to store the data.
con = create_context(host=host, user=user, password=password)
execute_sql("CREATE TABLE read_nos_support_tab (payload dataset storage format csv) NO PRIMARY INDEX;")
read_nos_support_tab = DataFrame("read_nos_support_tab")
# Read the CSV data using "data" argument.
obj = ReadNOS(data=read_nos_support_tab,
location="/S3/s3.amazonaws.com/Your-Bucket/csv-location/",
access_id="YOUR-ID",
access_key="YOUR-KEY",
row_format={"field_delimiter": ",", "record_delimiter": "\n", "character_set": "LATIN"}
stored_as="TEXTFILE")
# print the result DataFame.
print(obj.result)
# Note:
# Before proceeding, verify with your database administrator that you have the
# correct privileges, an authorization object, and a function mapping (for READ_NOS
# function).
# If function mapping for READ_NOS Analytic database function is created
# as 'READ_NOS_FM' and location and authorization object are mapped,
# then set function mapping with teradataml options as below.
# Example 5: Setting function mapping using configuration options.
from teradataml.options.configure import configure
configure.read_nos_function_mapping = "READ_NOS_FM"
obj = ReadNOS(data=read_nos_support_tab,
row_format={"field_delimiter": ",", "record_delimiter": "\n", "character_set": "LATIN"}
stored_as="TEXTFILE")
# print the result DataFame.
print(obj.result)
|