How to Properly Convert Multiple Column Classes in R data.table

---

If anything seems off to you, please feel free to write me at vlogize [AT] gmail [DOT] com.
---

Background: The Challenge at Hand

Born_before_2016

gender

payor

Age_in_day

Our goal is to change the classes of Born_before_2016, gender, and payor to character while leaving Age_in_day untouched. However, when performing this conversion, we noticed that Age_in_day was also being converted, which was not intended.

[[See Video to Reveal this Text or Code Snippet]]

The primary issue stems from how we identified which columns to convert and how we managed the selection of these columns in the .SDcols argument. Let’s dive into the steps to resolve this issue.

Step-by-Step Solution: Fixing the Column Conversion

1. Identify Columns to Convert

First, we need to specify which columns we want to convert. We can create a vector of those column names:

[[See Video to Reveal this Text or Code Snippet]]

Next, we use the sapply function to check which of these selected columns are numeric:

[[See Video to Reveal this Text or Code Snippet]]

2. Select Valid Column Names

After identifying numeric columns, we create a vector of the actual column names to pass to .SDcols. Note that factcols returns a logical vector, representing the column positions instead of their names.

[[See Video to Reveal this Text or Code Snippet]]

3. Perform the Conversion

Finally, we can perform the conversion from numeric to character. Here’s the corrected code that ensures we only convert the intended columns:

[[See Video to Reveal this Text or Code Snippet]]

Result Verification

When we execute the code, we can verify that the intended columns have been converted to character while Age_in_day remains unaffected:

[[See Video to Reveal this Text or Code Snippet]]

Now the Born_before_2016, gender, and payor columns are successfully converted to character strings, and Age_in_day remains as integer type as intended.

Conclusion