Data360 Analyze

Hi,

In lavastorm I can "fill in" the date range given 2 dates, outputing each date on a new line (See brainscript below)

j = 0

output 1 {emit FillInMonths}

while j < i
{
FillInMonths=FillInMonths.dateAdjust(1,"months")
do output 1 {emit FillInMonths}

j=j+1
}

Can I achieve the same thing in Python and how does the code look?  I have tried multiple options with no success.

thanks

regards

Doug

just to caveat, I'm trying to do this within in Transform node in Dataverse.

Hi Doug,

it is a little more convoluted in Python as the core Python does not provide a function to adjust a datetime object by a month (as it contains a variable number of days). Here is a solution that is taken from a post on StackOverflow:

#### This code outputs multiple records with an incrementing date for each input record

#### ConfigureFields Script

## Credit for custom Python function:
## https://stackoverflow.com/questions/35066588/is-there-a-simple-way-to-increment-a-datetime-object-one-month-in-python/35067328

import calendar

def add_one_month(orig_date):
  # advance year and month by one month
  new_year = orig_date.year
  new_month = orig_date.month + 1
  # note: in datetime.date, months go from 1 to 12
  if new_month > 12:
    new_year += 1
    new_month -= 12
 
  last_day_of_month = calendar.monthrange(new_year, new_month)[1]
  new_day = min(orig_date.day, last_day_of_month)
 
  return orig_date.replace(year=new_year, month=new_month, day=new_day)

######

#Configure all fields from input 'in1' to be mapped
#to the corresponding fields on the output 'out1'
out1 += in1

_FillInMonths = datetime.datetime(2015,2,28).date()

## Specify the number of records to be output
## for each input record (zero-based)
i = 20

## Define the new date field to hold the adjusted records
out1.FillInMonths = datetime.date

#### ProcessRecords Script

#Copy all fields from input 'in1' to the corresponding output fields
#in output 'out1'. Copies the values of the fields that have been setup
#in the mapping defined in the ConfigureFields property
out1 += in1

## Set the first date value
out1.FillInMonths = _FillInMonths
##explicitly write the output record
node.write(0, out1)

## Now loop around outputing an adjusted date
## in each record
j = 0

while j < i:
  _FillInMonths = add_one_month(_FillInMonths)
  out1.FillInMonths = _FillInMonths
  node.write(0, out1)
  j += 1

Thanks Adrian,

that works nicely.  I think my main problem has been figuring out how the transform node is expecting the python to be presented without it erroring.

I think your answer will help me move forward with other aspects of phython coding in Dataverse as well so thanks.

Doug

Just to be careful, what do you mean by "Add a month"?

January 31 + 30 days => March 2nd

August 31 + "1 month" will throw an error because it tries to create Sept 31.

The code mentioned above works for DayOfMonth < 29, but has troubles for days 29,30,31

I agree with Stony's warning about the interpretation of adding a month. However, I just tried it with an initial date of 2015-08-31 (so that the sequence will go through the leap day in Feb 2016). The following line in the code:

new_day = min(orig_date.day, last_day_of_month)

causes the new date to be pegged to the last day in the new month if it is less than the value for the adjusted day of month. This means the dates will 'degrade' from 31st through 30th down to (29th or) 28th of the month as the adjusted month tracks though the year.

Here is a modified version of the code that will increment the date and peg the date to the last day of the new month on a month by month basis when the starting day of month for the range is greater than 28. So, if you were trying to create a range of payment dates the specified starting day of month would be used where possible and where not, the last day of the month would be used.

#### ConfigureFields Script

## Credit for custom Python function:
## https://stackoverflow.com/questions/35066588/is-there-a-simple-way-to-increment-a-datetime-object-one-month-in-python/35067328
##
## Modified to allow the number of months to be optionally specified in the function call

import calendar

def add_n_months(orig_date, num_mths = 1):
  if not isinstance(num_mths, (int, long)):
    node.logger.error("The number of months must be an integer.")
    raise node.fail()
  if num_mths < 1:
    node.logger.error("The number of months must be at least 1.")
    raise node.fail()
  if num_mths > 11:
    years_to_add = num_mths/ 12 ## Integer division used by default
    months_to_add = num_mths % 12
  else:
    years_to_add = 0
    months_to_add = num_mths
  # advance year and month by number of months
  new_year = orig_date.year + years_to_add
  new_month = orig_date.month + months_to_add
  # note: in datetime.date, months go from 1 to 12
  if new_month > 12:
    new_year += 1
    new_month -= 12
 
  last_day_of_month = calendar.monthrange(new_year, new_month)[1]
  new_day = min(orig_date.day, last_day_of_month)
 
  return orig_date.replace(year=new_year, month=new_month, day=new_day)

######

#Configure all fields from input 'in1' to be mapped
#to the corresponding fields on the output 'out1'
out1 += in1

_startMonth = datetime.datetime(2015,8,31).date()

## Specify the number of records to be output
## for each input record (zero-based)
i = 20

## Define the new date field to hold the adjusted records
out1.FillInMonths = datetime.date

#### ProcessRecords Script

#Copy all fields from input 'in1' to the corresponding output fields
#in output 'out1'. Copies the values of the fields that have been setup
#in the mapping defined in the ConfigureFields property
out1 += in1

## Set the first date value
out1.FillInMonths = _startMonth
##explicitly write the output record
node.write(0, out1)

## Now loop around outputing an adjusted date
## in each record
j = 0

while j < i:
  j += 1
  _FillInMonths = add_n_months(_startMonth, j)
  out1.FillInMonths = _FillInMonths
  node.write(0, out1)