Ufdjrw

I want to make a dictionary with a specific set of keys and a specific set of values. These are the keys I want to use:

k = ['date', 'deviceCategory', 'transactionId', 'productSku', 'productName', 'productCategoryHierarchy', 'channelGrouping', 'itemRevenue', 'itemQuantity']

And these are the values for each key:

v = [datetime.date(2019, 3, 5), 'desktop', 1551740677701, 60104621, '(not set)', 'sale/apartment/alicante/bajo-vinalopo/elx', 'Tráfico de Búsqueda de Pago - Venta', 0.0, 1]

If I use a set or try to make a dictionary using this:

d = dict(zip(k, v))

Or even this (what I really intend to do):

d = dict(zip(map(lambda x: "ga_%s" % x, k), v))

I get this:

'ga_itemRevenue': 0.0, 'ga_itemQuantity': 1, 'ga_deviceCategory': 'desktop', 'ga_date': datetime.date(2019, 3, 5), 'ga_channelGrouping': 'Tráfico de Búsqueda de Pago - Venta', 'ga_productSku': 60104621, 'ga_productName': '(not set)', 'ga_productCategoryHierarchy': 'sale/apartment/alicante/bajo-vinalopo/elx', 'ga_transactionId': 1551740677701

I know key order is not relevant in python dictionaries (at least, the ones made with dict()) but I need the keys and their matching values to form the dictionary in the exact same order, which is not alphabetical. Just... the same key order.

I've tried making an ordered dictionary with collections.OrderedDict() and the order is the one I want except for the fact it gives me a list of tuples. I've also tried to kind of 'convert' the ordered dictionary into a regular one with this:

d = json.loads(json.dumps(x))

But it gives me an error because the datetime object cannot be serialized (or so it says the error message) and I need it like that (as a datetime object) as this is going to be written to a database table with the cx_Oracle package.

I've also read about frozenset() but seems like I need a previously existing set to 'freeze' it and once I create a set is exactly like trying to make a dictionary. keys are sprayed over the set and I need them in the same order they're declared in the keys array.

How can I achieve this?

#EDIT 1. I forgot to say this is done in Python 2.7.x

#EDIT 2. Here is the code fragment I'm having problems with:

reports = response.get("reports", [])

if len(reports) > 0:
 for report in reports:
 rows = report.get("data", ).get("rows", [])

 if len(rows) > 0:
 k = ga_dimensions + ga_metrics
 o = []

 for row in rows:
 o.append(map(lambda x, y: cast_field_type(x, y), row.get("dimensions", []) + row.get("metrics", [])[0]["values"], k))

 if len(o) > 0:
 # insert all data rows into the table
 for v in o:
 for i in range(0, len(v)):
 if k[i] == "date" and is_date(v[i]):
 v[i] = date(*map(int, v[i].split("-")))
 elif isinstance(v[i], unicode):
 v[i] = v[i].encode("utf-8")

 v = dict(zip(map(lambda x: "ga_%s" % x, k), v))
 cr.execute(q, v)

 # commit all changes
 db.commit()

This script fetches data from Google Analytics (using the v4 API), infer the appropiate data types (for database storage with Oracle) from the first row of data, traverse the data rows so specific fields (dates and Unicode strings, basically) get appropiately casted/converted/encoded before being sent to the database and, when done that, commit the changes so the data gets written for real.

I try to do this because the script gives me this error when trying to supply the dictionary per se:

cx_Oracle.DatabaseError: ORA-00932: inconsistent datatypes: expected DATE got NUMBER

And this is because the data is being supplied with a key order that is different from the SQL INSERT query being specified.

collections.OrderedDict seems to do what you want. To illustrate what @chepner mentioned about the representation of OrderedDict being tuples, I'll show you what I ran:

#encoding: utf-8
from collections import OrderedDict
import datetime

k = ['date', 'deviceCategory', 'transactionId', 'productSku', 'productName', 'productCategoryHierarchy', 'channelGrouping', 'itemRevenue', 'itemQuantity']
v = [datetime.date(2019, 3, 5), 'desktop', 1551740677701, 60104621, '(not set)', 'sale/apartment/alicante/bajo-vinalopo/elx', 'Tráfico de Búsqueda de Pago - Venta', 0.0, 1]

d = OrderedDict(zip(k,v))
for i in d:
 print(': '.format(i,d[i]))

print('nn')

print(d)

and this was the output:

date: 2019-03-05
deviceCategory: desktop
transactionId: 1551740677701
productSku: 60104621
productName: (not set)
productCategoryHierarchy: sale/apartment/alicante/bajo-vinalopo/elx
channelGrouping: Tráfico de Búsqueda de Pago - Venta
itemRevenue: 0.0
itemQuantity: 1



OrderedDict([('date', datetime.date(2019, 3, 5)), ('deviceCategory', 'desktop'), ('transactionId', 1551740677701), ('productSku', 60104621), ('productName', '(not set)'), ('productCategoryHierarchy', 'sale/apartment/alicante/bajo-vinalopo/elx'), ('channelGrouping', 'Trxc3xa1fico de Bxc3xbasqueda de Pago - Venta'), ('itemRevenue', 0.0), ('itemQuantity', 1)])

So it looks like it's behaving as you were trying to get it to.

The above solution might be what you need, but if you used pandas you'd get an intermediate dataframe which makes manipulating, analyzing, and visualizing data easier (if that's an avenue you wanted). The solution looks something like this:

import datetime
import pandas as pd

k = ['date', 'deviceCategory', 'transactionId', 'productSku', 'productName', 'productCategoryHierarchy', 'channelGrouping', 'itemRevenue', 'itemQuantity']

v = [datetime.date(2019, 3, 5), 'desktop', 1551740677701, 60104621, '(not set)', 'sale/apartment/alicante/bajo-vinalopo/elx', 'Tráfico de Búsqueda de Pago - Venta', 0.0, 1]

df = pd.concat([pd.DataFrame(v, k)], axis=1)

# the dataframe
date 2019-03-05
deviceCategory desktop
transactionId 1551740677701
productSku 60104621
productName (not set)
productCategoryHierarchy sale/apartment/alicante/bajo-vinalopo/elx
channelGrouping Tráfico de Búsqueda de Pago - Venta
itemRevenue 0
itemQuantity 1

dict = df.to_dict()
results = dict[0]
results

# the dictionary
'date': datetime.date(2019, 3, 5),
 'deviceCategory': 'desktop',
 'transactionId': 1551740677701,
 'productSku': 60104621,
 'productName': '(not set)',
 'productCategoryHierarchy': 'sale/apartment/alicante/bajo-vinalopo/elx',
 'channelGrouping': 'Tráfico de Búsqueda de Pago - Venta',
 'itemRevenue': 0.0,
 'itemQuantity': 1

Two important things to note:

• make sure you use bind variables (for performance and security).
  This will also let you name the variables so the data order shouldn't
  matter


• when inserting lots of rows it is MUCH more efficient to
  use executeMany() as discussed in
  https://blogs.oracle.com/opal/efficient-and-scalable-batch-statement-execution-in-python-cx_oracle