P2D5: Writing Python code

Using pd.read_fwf()

pd.read_fwf() doesn’t provide much detail on how to use the function.

1. Let’s start with one of our files
2. What arguments do we have available to us in pd.read_fwf().
   • path_or_buf
   • colspecs
   • widths
   • **kwds What are these? Kew word arguments from pd.read_table()
3. Ok, now we have one dataset locally and we have it parsed into Python. Let’s write psuedo code for our Python script.

range objects and list comprehension

In R we have access to vectors of numbers using :. It is intuitive to me and easy to read in code.

range_variable <- 60:84

R also converts numbers to text automatically with paste.

paste0("X", range_variable)

Then we can add more items with c().

c(paste0("X", range_variable), "U03", "Y870")

Python is not as friendly to the R user mind. We have to understand a few programming constructs to create a similar variable.

1. range() objects.
2. Convert numbers to strings with str()
3. list comprehension
4. joining lists

# My python code to make the magic happen??
range(60, 85) # to start, but why 85?

Creating the suicide table in Pandas

Here is the example R code.

suicide <- raw_file %>%
    dplyr::filter(underlying_cause %in% suicide_code) %>%
    dplyr::mutate(
      gun = ifelse(underlying_cause %in% c("X72", "X73", "X74"), 1, 0),
      year = year
      )

• .query()
• .assign()
• .isin()
• np.where()

Creating the guns table in Pandas

Pandas does not have a clean case_when() like dplyr. We can use np.select() to get close.

intent_cond = [
    (raw_file.underlying_cause.isin(["W32", "W33", "W34"])),
    (raw_file.underlying_cause.isin(["X72", "X73", "X74"])),
    (raw_file.underlying_cause.isin(["*U01.4", "X93", "X94", "X95",
          "Y350"])),
    (raw_file.underlying_cause.isin(["Y22", "Y23", "Y24"]))
]

intent_val = ["Accidental", "Suicide", "Homicide", "Undetermined"]

guns = raw_file.assign(
        intent = np.select(intent_cond, intent_val, default = np.nan)
    )