Skip to content

Manage column profiling information

Profiling gives additional context to columns in relational stores. From profiling, you can see various summarized information such as:

  • numerical statistics (min, max, mean, median, standard deviation, sum, variance) for numeric columns
  • minimum, maximum, and average lengths for string columns
  • distinct value counts and percentages
  • missing value counts and percentages

Profiling is only available on columns

You will only be able to populate this summary information on columns, not on other assets in Atlan.

Retrieve profiles

1.4.0 1.0.0

Since profiles are only available on columns, you will need to retrieve column assets to see the profiles:

Retrieve profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
Column column = Column.get( // (1)
        "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS"); // (2)
column.getColumnDistinctValuesCount(); // (3)
column.getColumnUniqueValuesCount();
column.getColumnUniquenessPercentage();
column.getColumnDuplicateValuesCount();
column.getColumnMissingValuesCount();
column.getColumnMissingValuesPercentage();
column.getColumnMax(); // (4)
column.getColumnMin();
column.getColumnMean();
column.getColumnMedian();
column.getColumnStandardDeviation();
column.getColumnVariance();
column.getColumnSum();
column.getColumnMinimumStringLength(); // (5)
column.getColumnMaximumStringLength();
column.getColumnAverageLength();
  1. Use the get() method to retrieve all details about a specific column.
  2. Provide the full, case-sensitive qualifiedName of the column.
  3. Some profile information is common, regardless of the data type of the column.
  4. Some profile information is specific to numeric columns.
  5. Some profile information is specific to string columns.
Retrieve profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
from pyatlan.client.atlan import AtlanClient
from pyatlan.model.assets import Column

client = AtlanClient()
column = client.asset.get_by_qualified_name(  # (1)
    qualified_name="default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS",  # (2)
    asset_type=Column
)
column.column_distinct_values_count  # (3)
column.column_unique_values_count
column.column_uniqueness_percentage
column.column_duplicate_values_count
column.column_missing_values_count
column.column_missing_values_percentage
column.column_max  # (4)
column.column_min
column.column_mean
column.column_median
column.column_standard_deviation
column.column_variance
column.column_sum
column.column_minimum_string_length  # (5)
column.column_maximum_string_length
column.column_average_length
  1. Use the get_by_qualified_name() method to retrieve all details about a specific column.
  2. Provide the full, case-sensitive qualifiedName of the column.
  3. Some profile information is common, regardless of the data type of the column.
  4. Some profile information is specific to numeric columns.
  5. Some profile information is specific to string columns.
Retrieve profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
val column = Column.get( // (1)
    "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS") // (2)
column.columnDistinctValuesCount // (3)
column.columnUniqueValuesCount
column.columnUniquenessPercentage
column.columnDuplicateValuesCount
column.columnMissingValuesCount
column.columnMissingValuesPercentage
column.columnMax // (4)
column.columnMin
column.columnMean
column.columnMedian
column.columnStandardDeviation
column.columnVariance
column.columnSum
column.columnMinimumStringLength // (5)
column.columnMaximumStringLength
column.columnAverageLength
  1. Use the get() method to retrieve all details about a specific column.
  2. Provide the full, case-sensitive qualifiedName of the column.
  3. Some profile information is common, regardless of the data type of the column.
  4. Some profile information is specific to numeric columns.
  5. Some profile information is specific to string columns.
GET /api/meta/entity/uniqueAttribute/type/Column?attr:qualifiedName=default%2Fhive%2F1657025257%2FOPS%2FDEFAULT%2FRUN_STATS%2FSTATUS
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
{
  "entity": { // (1)
    "typeName": "Column", // (2)
    "attributes": { // (3)
      "name": "STATUS",
      "qualifiedName": "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS",
      "columnDistinctValuesCount": 123, // (4)
      "columnUniqueValuesCount": 123,
      "columnUniquenessPercentage": 50.0,
      "columnDuplicateValuesCount": 123,
      "columnMissingValuesCount": 123,
      "columnMissingValuesPercentage": 50.0,
      "columnMax": 321.0,
      "columnMin": 1.0,
      "columnMean": 123.0,
      "columnMedian": 123.0,
      "columnStandardDeviation": 3.0,
      "columnVariance": 1.0,
      "columnSum": 654321.0,
      "columnMinimumStringLength": 0,
      "columnMaximumStringLength": 123,
      "columnAverageLength": 123.0
    }
  }
}
  1. All column details will come back wrapped in a top-level entity object in the payload.
  2. The typeName will always be Column.
  3. The detailed profiling information will be embedded in the attributes object within the outer entity object.
  4. The column profiling details have names that all start with column...

Add your own profiles

1.4.0 1.0.0

In cases where Atlan does not profile the source, you may want to add your own profiles. You can do this by either adding the profile when creating the column (programmatically) or by updating the attributes on an existing column:

Add or update profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
Column column = Column.updater( // (1)
        "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS", // (2)
        "STATUS") // (3)
    .columnDistinctValuesCount(123) // (4)
    .columnUniqueValuesCount(123)
    .columnUniquenessPercentage(50.0)
    .columnDuplicateValuesCount(123)
    .columnMissingValuesCount(123)
    .columnMissingValuesPercentage(50.0)
    .columnMax(321.0) // (5)
    .columnMin(1.0)
    .columnMean(123.0)
    .columnMedian(123.0)
    .columnStandardDeviation(3.0)
    .columnVariance(1.0)
    .columnSum(654321.0)
    .columnMinimumStringLength(0) // (6)
    .columnMaximumStringLength(123)
    .columnAverageLength(123.0)
    .build(); // (7)
AssetMutationResponse response = column.save(); // (8)
  1. Use the updater() method to update an existing column asset (for more details, see Updating an asset).
  2. Provide the full, case-sensitive qualifiedName of the column.
  3. Provide the case-sensitive name of the column.
  4. Some profile information is common, regardless of the data type of the column. All are optional, so fill in only the pieces you want or for which you have the information.
  5. Some profile information is specific to numeric columns. All are optional, so fill in only the pieces you want or for which you have the information.
  6. Some profile information is specific to string columns. All are optional, so fill in only the pieces you want or for which you have the information.
  7. Use the build() method to construct the column object to be updated in Atlan.
  8. Then call the save() method against this built-up object to actually apply the update to Atlan.
Retrieve profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
from pyatlan.client.atlan import AtlanClient
from pyatlan.model.assets import Column

client = AtlanClient()
column = Column.create_for_modification(  # (1)
    qualified_name="default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS",  # (2)
    name="STATUS"  # (3)
)
column.column_distinct_values_count = 123  # (4)
column.column_unique_values_count = 123
column.column_uniqueness_percentage = 50.0
column.column_duplicate_values_count = 123
column.column_missing_values_count = 123
column.column_missing_values_percentage = 50.0
column.column_max = 321.0  # (5)
column.column_min = 1.0
column.column_mean = 123.0
column.column_median = 123.0
column.column_standard_deviation = 3.0
column.column_variance = 1.0
column.column_sum = 654321.0
column.column_minimum_string_length = 0  # (6)
column.column_maximum_string_length = 123
column.column_average_length = 123.0
response = client.asset.save(column)  # (7)
  1. Use the create_for_modification() method to update an existing column asset (for more details, see Updating an asset).
  2. Provide the full, case-sensitive qualified_name of the column.
  3. Provide the case-sensitive name of the column.
  4. Some profile information is common, regardless of the data type of the column. All are optional, so fill in only the pieces you want or for which you have the information.
  5. Some profile information is specific to numeric columns. All are optional, so fill in only the pieces you want or for which you have the information.
  6. Some profile information is specific to string columns. All are optional, so fill in only the pieces you want or for which you have the information.
  7. Then call the save() method with this built-up object to actually apply the update to Atlan.
Add or update profiles
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
val column = Column.updater( // (1)
        "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS",  // (2)
        "STATUS") // (3)
    .columnDistinctValuesCount(123) // (4)
    .columnUniqueValuesCount(123)
    .columnUniquenessPercentage(50.0)
    .columnDuplicateValuesCount(123)
    .columnMissingValuesCount(123)
    .columnMissingValuesPercentage(50.0)
    .columnMax(321.0) // (5)
    .columnMin(1.0)
    .columnMean(123.0)
    .columnMedian(123.0)
    .columnStandardDeviation(3.0)
    .columnVariance(1.0)
    .columnSum(654321.0)
    .columnMinimumStringLength(0) // (6)
    .columnMaximumStringLength(123)
    .columnAverageLength(123.0)
    .build() // (7)
val response = column.save() // (8)
  1. Use the updater() method to update an existing column asset (for more details, see Updating an asset).
  2. Provide the full, case-sensitive qualifiedName of the column.
  3. Provide the case-sensitive name of the column.
  4. Some profile information is common, regardless of the data type of the column. All are optional, so fill in only the pieces you want or for which you have the information.
  5. Some profile information is specific to numeric columns. All are optional, so fill in only the pieces you want or for which you have the information.
  6. Some profile information is specific to string columns. All are optional, so fill in only the pieces you want or for which you have the information.
  7. Use the build() method to construct the column object to be updated in Atlan.
  8. Then call the save() method against this built-up object to actually apply the update to Atlan.
POST /api/meta/entity/bulk
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
{
  "entities": [ // (1)
    {
      "typeName": "Column", // (2)
      "attributes": {
        "name": "STATUS", // (3)
        "qualifiedName": "default/hive/1657025257/OPS/DEFAULT/RUN_STATS/STATUS", // (4)
        "columnDistinctValuesCount": 123, // (5)
        "columnUniqueValuesCount": 123,
        "columnUniquenessPercentage": 50.0,
        "columnDuplicateValuesCount": 123,
        "columnMissingValuesCount": 123,
        "columnMissingValuesPercentage": 50.0,
        "columnMax": 321.0, // (6)
        "columnMin": 1.0,
        "columnMean": 123.0,
        "columnMedian": 123.0,
        "columnStandardDeviation": 3.0,
        "columnVariance": 1.0,
        "columnSum": 654321.0,
        "columnMinimumStringLength": 0, // (7)
        "columnMaximumStringLength": 123,
        "columnAverageLength": 123.0
      }
    }
  ]
}
  1. All columns must be wrapped in an entities array.
  2. The typeName must always be Column for profiling information.
  3. You must provide the exact name of the column (case-sensitive).
  4. You must provide the exact qualifiedName of the column (case-sensitive).
  5. Some profile information is common, regardless of the data type of the column. All are optional, so fill in only the pieces you want or for which you have the information.
  6. Some profile information is specific to numeric columns. All are optional, so fill in only the pieces you want or for which you have the information.
  7. Some profile information is specific to string columns. All are optional, so fill in only the pieces you want or for which you have the information.