Add_field adding values based on other column (aka conditional values)

Copying from Discord channel, with some updated info:

Hi there! As Framework Frictionless documentation is still progress, I would like to ask whether someone can advice me regarding how to write custom step function? Coming from dataflows, my goal is to translate this to Framework:

def flow(parameters, datapackage, resources, stats):

    def category():
        def step(row):
            if row['field_1'] is not None:
                if float(row['field_1']) > 10:
                    row['Category'] = 'A'
                else:
                    row['Category'] = 'B '
        return step

    return Flow(
        add_field('Category', 'string'),
        category()
    )

In Framework I would use:

source = Resource(path="resource.csv")
target = transform(
    source,
    steps=[
        steps.field_add(name="field_1", type="string", value="A/B"), # <--HERE
    ]
)

Basically, question is how to make adding conditional value (as in this example “A” or “B”) depending on other column value from resource.Thanks.

UPDATE: I came with this, however I am not able to use comparison: if rec[‘field_1’] < 11 rather than if (rec[‘field_1’] in str(list(range(11))))

steps=[steps.field_add(name="Category",
                                     type='string', 
                                     value=lambda rec: 'A' if (rec['field_1'] in str(list(range(11)))) else 'B'),
                                     ]

[roll]

Hi,

there are two options.

You can cast the data in the function:

source = Resource(path="inp.csv")
target = transform(
source,
steps=[
    steps.field_add(name="Category", type='string', function=lambda rec: '<100' if int(rec['BTOT'])>100 else '1-100'),])

Or you can normalize everything beforehand (less effective on large data):

source = Resource(path="inp.csv")
target = transform(
source,
steps=[
    spets.table_normalize(),
    steps.field_add(name="Category", type='string', function=lambda rec: '<100' if rec['BTOT']>100 else '1-100'),])

[roll]

Please re-open if you need more help 👍