Delete duplicates based on varchar in another row. Eliminate terminated staff from active roster - drop-duplicates

Question

I want to create a roster that shows only active staff. All staff are listed in a table, active staff with no termination have one row in the table. Terminated staff have two rows of data. Employee status is wages or term. How can I get one row of only active staff? Each staff member has a unique ID.

Answer 1

What programming language are you using?
I'm assuming you're using Python, and using pandas to process data.
Also, whenever possible, sharing a sample data enables other users
to help you more efficiently.
Please, if possible, edit your question adding a little more context to your question, so that others can help you more efficiently!
I've developed the code below, trying to re-create your problem based on what I could infer about the data from your question.
# == Necessary Imports ===============================================
import pandas as pd
# == Generate sample dataframe =======================================
# For our sample, employees 3, 7, and 9 will represent the terminated
# employees.
df = pd.DataFrame(
{
"memberID": [1, 2, 3, 3, 4, 5, 6, 7, 7, 8, 9, 9],
"status": [
"wages",
"wages",
"wages", # <-- Employee 3 terminated
"term", # <-- Employee 3 terminated
"wages",
"wages",
"wages",
"wages", # <-- Employee 7 terminated
"term", # <-- Employee 7 terminated
"wages",
"wages", # <-- Employee 9 terminated
"term", # <-- Employee 9 terminated
],
}
)
# == Solution ==============================================================
df.loc[
df["memberID"].isin(
df["memberID"]
.value_counts()
.to_frame("count")
.loc[lambda xdf: xdf["count"] < 2, :]
.index
),
:,
]
# Returns:
#
# memberID status
# 0 1 wages
# 1 2 wages
# 4 4 wages
# 5 5 wages
# 6 6 wages
# 9 8 wages
Notes
Basically, my approach was to use pandas.Series.value_counts to find the number of times each "memberID" appears on the dataframe.
Then I chained a .loc call to filter for employees with less than 2 appearances on the dataframe and filtered those "memberID" on the original df.
In summary, the solution uses three methods from pandas:
pandas.Series.value_counts: count the number of times each value from a pandas Series gets repeated.
pandas.Series.to_frame: transform a pandas Series to DataFrame (useful for chaining multiple operations together).
pandas.DataFrame.loc: used to filter employees that appeared only once in the data. Then used a second time to filter the original df to only consider those employees that weren't terminated.