Passing Parameters/Variables into python Script

[Anonymous]

Dear Community,

I am trying to build a function for web scrapping using a python script. I want to pass a parameter as a variable to the python script.  I tried following the parameter passing formula in R script. But that is not working for python file. 

Is there a way to declare the search 'query' as a parameter and have it passed through or stored as a variable to be used in each time the function is invoked?

I have also shared my code snippet for reference. 

Thanks.

[Anonymous]

You need to concatenate text with & and not +

The following works fine for me:

let
Source = Python.Execute("import pandas#(lf) azureAppid='" & paramAzureAppID & "'#(lf)",

df1 = Source{[Name="df"]}[Value]
in
df1

[v-danhe-msft]

Hi @Anonymous ,

Could you have refered to below link about Using Python in Query Editor:

https://docs.microsoft.com/en-us/power-bi/desktop-python-in-query-editor

Maybe you could just enter put the python code into the Run Python Script dialog.

Regards,
Daniel He

[Anonymous]

Hi @v-danhe-msft 

Thank you for your reply. 

I was trying to build a function with the python script. The parameter passing is not working for the function. 

But the python script is working fine when it was added as a new Step. 

I was wondering if there is any way I could make the same script as a function.

Regards,

Jannatun Nahar

[Anonymous]

Hey, @Anonymous 

Can you explain more about "passing as a new step"? Were you able to find a workaround to use the parameter as part of your Python script?

Thanks!

[Anonymous]

Hi @Anonymous ,

Thank you very much for your reply. I am trying to create a custom invoked function using python or R script to extract tweets in powerBI. Instead of hardcoding the search term in the original R script, I wanted to pass it as a parameter. But the code is not working in this way. I am attaching a copy of my code here :

(Name as text) as table=>
let
    
  Source = R.Execute(" #(lf)   library(twitteR)
                         #(lf)   library(stringi)
                         #(lf)   consumer_key <- ""abcd "" 
                         #(lf)   consumer_secret <- "" adbc ""
                         #(lf)   access_token <- ""abcd ""
                         #(lf)   access_secret <- ""abcd""
                         #(lf)   setup_twitter_oauth(consumer_key, consumer_secret, access_token, access_secret)
                         #(lf)   str1 <- ""#""
                         #(lf)   str2 <- ""&Name&""   (passing serach term as a parameter here)
                         #(lf)   search_term <- paste(str1,str2,sep="") 
                         #(lf)   number_of_tweets <- 100
                         #(lf)   language <-'en'
                         #(lf)   tweets <- searchTwitter(search_term, number_of_tweets,language)
                         #(lf)   df_tweets <- twListToDF(tweets)
                         #(lf)   df_tweets <- df_tweets[,c('text','favorited','favoriteCount','created','retweetCount',
                                                                                              'isRetweet','retweeted')]
                         #(lf)   df_tweets$text <- stri_encode(df_tweets$text, '', 'UTF-8')"),
    df_tweets1 = Source{[Name="df_tweets"]}[Value]
in
    df_tweets1

I am working on this for my MSC capstone project. I would really appreciate any suggestion to pass through the error. 

Thanks agaian.

[deep_thinker]

Can someone please confirm whether this is possible or not? 

[armanschwarz]

This seems to work - note that dataset2 takes the dataset1 value and multiplies it by 2:

let
Source = Python.Execute("import pandas#(lf)dataset = pandas.DataFrame([1, 2, 3])"),
dataset1 = Source{[Name="dataset"]}[Value],
dataset2 = Python.Execute("
import pandas
dataset = dataset*2
dataset.to_csv('C:/Users/arman.schwarz/dataset_to_csv.csv')", [dataset = dataset1]),
dataset3 = dataset2{[Name="dataset"]}[Value],
#"Changed Type" = Table.TransformColumnTypes(dataset3,{{"0", Int64.Type}})
in
#"Changed Type"