code snippets for doing tidyverse things in pandas

pandas_tidyverse_equivalents.md

Code for wide to long to wide again

• Python MultiIndex artifacts mean that df.pivot() is not the exact inverse of df.melt()

# ---- "load" wide data ----

# df <- tibble::tribble(~country, ~pop, ~capital,
#               "USA",      330,  "Washington D.C.",
#               "AUS",      25,   "Canberra",
#               "JPN",      125,  "Tokyo" )

df = pd.DataFrame({'country': ['USA', 'AUS', 'JPN'],
                   'pop': [330, 25, 125],
                   'capital': ['Washington D.C.', 'Canberra', 'Tokyo']})

print("--- wide data ---\n", df)

# ---- wide to long format ----

# df_long <- df %>% 
#      tidyr::pivot_longer(names_to = "variable", values_to = "value",
#                          cols = -c(country))

df_long = df.melt(id_vars=['country'])
print("--- long format ---\n", df_long)

# ---- back to wide ----

# df_long <- df %>%
#      pivot_wider(names_from = "variable", values_from = "value")

df_wide = df_long.pivot(index=['country'], 
                   columns='variable', values='value')
print("--- back to wide? ---\n", df_wide)

print("--- actually back to wide! ---")
(df_wide.reset_index()             ## put index (country) in a column
        .rename_axis(None, axis=1) ## removes column axis label 'variable'
).pipe(print)

Output

--- wide data ---
   country  pop          capital
0     USA  330  Washington D.C.
1     AUS   25         Canberra
2     JPN  125            Tokyo
--- long format ---
   country variable            value
0     USA      pop              330
1     AUS      pop               25
2     JPN      pop              125
3     USA  capital  Washington D.C.
4     AUS  capital         Canberra
5     JPN  capital            Tokyo
--- back to wide? ---
 variable          capital  pop
country                       
AUS              Canberra   25
JPN                 Tokyo  125
USA       Washington D.C.  330
--- actually back to wide! ---
  country          capital  pop
0     AUS         Canberra   25
1     JPN            Tokyo  125
2     USA  Washington D.C.  330