player & match df has match ids in an unusual format

[Dazzalytics]

First up, great work in making this package.

I was working with PSL data and noticed that player & match datasets have match ids as a long string of file paths. E.g. This long string C:\Users\AppData\Local\Temp\Rtmpuu2sV2/psl_male_bbb/1075986 shows up as match ids in data, where 1075986 is the match id.

I wrote this simple function below to fix it for myself. But probably there is a way to fix it in the package code. Thank you!

clean_match_id = function(df){ sapply(strsplit(df$match_id, "/"), '[',3) }

[robjhyndman]

I think this might depend on your operating system. Here's what I get on ubuntu:

library(cricketdata)
fetch_cricsheet(type = "match", gender = "male", competition = "psl")
#> # A tibble: 206 × 25
#>    match_id balls_per_over team1    team2 gender season date  event match_number
#>    <chr>    <chr>          <chr>    <chr> <chr>  <chr>  <chr> <chr> <chr>       
#>  1 1075986  6              Islamab… Pesh… male   2016/… 2017… Paki… 1           
#>  2 1075988  6              Karachi… Pesh… male   2016/… 2017… Paki… 3           
#>  3 1075995  6              Islamab… Kara… male   2016/… 2017… Paki… 10          
#>  4 1075997  6              Islamab… Pesh… male   2016/… 2017… Paki… 12          
#>  5 1076001  6              Lahore … Pesh… male   2016/… 2017… Paki… 16          
#>  6 1076005  6              Islamab… Kara… male   2016/… 2017… Paki… 20          
#>  7 1076007  6              Karachi… Isla… male   2016/… 2017… Paki… <NA>        
#>  8 1076008  6              Peshawa… Kara… male   2016/… 2017… Paki… <NA>        
#>  9 1075994  6              Peshawa… Quet… male   2016/… 2017… Paki… 9           
#> 10 1075990  6              Karachi… Quet… male   2016/… 2017… Paki… 5           
#> # … with 196 more rows, and 16 more variables: venue <chr>, city <chr>,
#> #   toss_winner <chr>, toss_decision <chr>, player_of_match <chr>,
#> #   umpire1 <chr>, umpire2 <chr>, reserve_umpire <chr>, tv_umpire <chr>,
#> #   match_referee <chr>, winner <chr>, winner_wickets <chr>, method <chr>,
#> #   winner_runs <chr>, outcome <chr>, eliminator <chr>
fetch_cricsheet(type = "player", gender = "male", competition = "psl")
#> # A tibble: 4,534 × 3
#>    team             player        match_id
#>    <chr>            <chr>         <chr>   
#>  1 Islamabad United DR Smith      1075986 
#>  2 Islamabad United Sharjeel Khan 1075986 
#>  3 Islamabad United BJ Haddin     1075986 
#>  4 Islamabad United SR Watson     1075986 
#>  5 Islamabad United SW Billings   1075986 
#>  6 Islamabad United Misbah-ul-Haq 1075986 
#>  7 Islamabad United Imran Khalid  1075986 
#>  8 Islamabad United Amad Butt     1075986 
#>  9 Islamabad United Saeed Ajmal   1075986 
#> 10 Islamabad United Mohammad Sami 1075986 
#> # … with 4,524 more rows

^{Created on 2022-02-20 by the reprex package (v2.0.1)}

I'll need to find a Windows computer to test it on.

[robjhyndman]

@jacquietran Are you using Windows? Can you replicate this issue?

[jacquietran]

Hey @robjhyndman and @Dazzalytics !

I can reproduce the issue on Windows:

> library(cricketdata)

> fetch_cricsheet(type = "match", gender = "male", competition = "psl")
# trying URL 'https://cricsheet.org/downloads/psl_male_csv2.zip'
# Content type 'application/zip' length 1002739 bytes (979 KB)
# downloaded 979 KB

# A tibble: 206 x 25                                                                                     
#   match_id     balls_per_over team1  team2  gender season date  event
#   <chr>        <chr>          <chr>  <chr>  <chr>  <chr>  <chr> <chr>
# 1 "C:\\Users\~ 6              Islam~ Pesha~ male   2016/~ 2017~ Paki~
# 2 "C:\\Users\~ 6              Karac~ Pesha~ male   2016/~ 2017~ Paki~
# 3 "C:\\Users\~ 6              Islam~ Karac~ male   2016/~ 2017~ Paki~
# 4 "C:\\Users\~ 6              Islam~ Pesha~ male   2016/~ 2017~ Paki~
# 5 "C:\\Users\~ 6              Lahor~ Pesha~ male   2016/~ 2017~ Paki~
# 6 "C:\\Users\~ 6              Islam~ Karac~ male   2016/~ 2017~ Paki~
# 7 "C:\\Users\~ 6              Karac~ Islam~ male   2016/~ 2017~ Paki~
# 8 "C:\\Users\~ 6              Pesha~ Karac~ male   2016/~ 2017~ Paki~
# 9 "C:\\Users\~ 6              Pesha~ Quett~ male   2016/~ 2017~ Paki~
# 10 "C:\\Users\~ 6              Karac~ Quett~ male   2016/~ 2017~ Paki~
# ... with 196 more rows, and 17 more variables: match_number <chr>,
#   venue <chr>, city <chr>, toss_winner <chr>, toss_decision <chr>,
#   player_of_match <chr>, umpire1 <chr>, umpire2 <chr>,
#   reserve_umpire <chr>, tv_umpire <chr>, match_referee <chr>,
#   winner <chr>, winner_wickets <chr>, method <chr>,
#   winner_runs <chr>, outcome <chr>, eliminator <chr>

> fetch_cricsheet(type = "player", gender = "male", competition = "psl")
# A tibble: 4,534 x 3                                                                                                 
#   team             player        match_id                                         
#   <chr>            <chr>         <chr>                                            
# 1 Islamabad United DR Smith      "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 2 Islamabad United Sharjeel Khan "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 3 Islamabad United BJ Haddin     "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 4 Islamabad United SR Watson     "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 5 Islamabad United SW Billings   "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 6 Islamabad United Misbah-ul-Haq "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 7 Islamabad United Imran Khalid  "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 8 Islamabad United Amad Butt     "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 9 Islamabad United Saeed Ajmal   "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# 10 Islamabad United Mohammad Sami "C:\\Users\\jacqu\\AppData\\Local\\Temp\\Rtmp29I~
# ... with 4,524 more rows

[robjhyndman]

It should be fixed now: https://github.com/robjhyndman/cricketdata/commit/80ded20df6446647a804a0b97fb0a9e5658becce

[Dazzalytics]

Thank you for the prompt feedback.

I have 0.1.1 version of the package and unfortunately, this issue is still there. I do not have package building experience, but I would like to help fix this issue (with some guidance). And possibly look at adding some more functionality to the package or work on ideas that you might already have.

[robjhyndman]

You will need to reinstall the package from github. Then try it. I have tested it on Windows and it worked for me on two computers.

[Dazzalytics]

Yes, it is working now. I appreciate your help!