Fix fill merged cells

[sjewo]

Hi @JanMarvin,

i found a minor problem while reading merged cells. If the merged area only contains NA values the import fails with an error. I saw this error in combination with startRow (after selecting a different startRow the merged cells became all NA)

library(openxlsx2)
library(testthat)

wb <- wb_workbook()
wb$add_worksheet("sheet1")
wb$add_data(1, t(matrix(c(1:3, NA_real_), 4, 4)), startRow = 3, startCol = 1, colNames = FALSE)

wb$merge_cells(1, rows = 1:4, cols = 4)

tmp_file <- temp_xlsx()
wb_save(wb, tmp_file)

# runs fine
read_xlsx(tmp_file, fillMergedCells = FALSE)

# Error in x[[jj]][iseq] <- vjj : replacement has length zero
read_xlsx(tmp_file, fillMergedCells = TRUE)

A simple solution is checking if the merged cells are all NA and skip this cell range. I created a pull request with a possible solution and a new test.

[JanMarvin]

Thanks, this might be related to #324. I never got to recreate the issue myself. Assumed that it was caused by merged cells overlapping into the requested region. Like every cell in row 1 is filled and a merged cell with column 2. Column 2 is empty, but the user requests to skip column 1 and fillMergedCells = TRUE.

Anyhow, nice catch and thanks for the PR :)