cr_extract: Extract a variable from a Cost Report
In medicare: Tools for Obtaining and Cleaning Medicare Public Use Files

Description Usage Arguments Details Value Examples

This function takes a 5-column alpha-numeric dataset or numeric dataset from the Medicare cost reports, which are stored in a long format, and subsets them based on the worksheet number, line number, and column number provided. If desired, it will rename the resulting variable to whatever the user chooses.

1	cr_extract(dataset, worksheet, row, column, newname = "newvar")

`dataset`	The name of a cost report alpha or numeric dataset
`worksheet`	The name of the workheet, converted to 7-character format
`row`	The row number of the data, as it appears in the Medicare workbook or documentation (i.e. at least 3 digits. Row 5 must be entered as 500, row 5.1 as 501, etc.)
`column`	The column number of the data, as it appears in the Medicare workbook (same general rule as for rows)
`newname`	The name given to the variable that appears as a result of this extraction

It does not automatically adjust for the same variable having different rows / columns in Medicare data formatted for the 1996 vs 2010 form. The user may have to use this function twice, once on each source of data, to extract one variable over time.

It does automatically recode rows and columns into all possible permutations (ie '500', '0500', '00500', 500) when subsetting, since different cost reports use different schema.

This function *does not* throw an error if the parameters yield an empty dataset at any point. It only gives warnings. This is because oftentimes the parameters are valid but the data is missing in the source material, due to CMS scrubbing of what data gets published.

A 2-column dataset: one with the cost report rpt_rec_number, used to merge data, and a column of the data requested, which is renamed if desired.