Skip to contents

Extract a rectangular block from a dataframe and make it to a matrix

Source: R/extract_data.R
extract_data.Rd

This function extracts a rectangular block from a dataframe using user-specified top/bottom row indices and left/right column identifiers (numeric or Excel-style letters). It ensures the block contains only numeric values or NAs, and returns a cleaned matrix.

Usage

extract_data(
  data,
  bottom_row,
  right_col,
  top_row = 1,
  left_col = 1,
  feature_name_columns = NA,
  use_row_index = FALSE
)

Arguments

data

data.frame: A dataframe containing the full input, including annotation columns and the numeric block to extract.

bottom_row

integer(1): Specifies the last (bottom) row of the numeric data block. Must be >= top_row.

right_col

integer(1) | character(1): Same format as left_col. Specifies the right-most column of the numeric block. Must be >= left_col after conversion.

top_row

integer(1): Specifies the first (top) row of the numeric data block. Row indexing is 1-based.

left_col

integer(1) | character(1): Column specifier for the left-most column of the data block. Can be either:

  • An integer index (e.g., 2 for the second column), or

  • A character string using Excel-style letters (e.g., "A", "AB").

Column names (e.g., "age") are not allowed here. Only letters or numeric indices are accepted.

feature_name_columns

character() | NULL: Optional character vector specifying columns in data to be used as row (feature) names in the output. If NA, generic feature names are used. These row names are used everywhere to label the features, such as to label the plots in the cluster_hits() function report.

use_row_index

logical(1): If TRUE, prepend the row index to each combined feature name to ensure uniqueness. Defaults to FALSE.

Value

A numeric matrix with cleaned values and appropriate column names.

Examples

# Tiny demo table with two header rows, feature columns, and numeric block
df <- data.frame(
    feat_id = c(NA, NA, "g1", "g2", "g3"),
    feat_sym = c(NA, NA, "TP53", "EGFR", "BAX"),
    A = c("cond", "t0", 1, 2, 3),
    B = c("cond", "t1", 4, 5, 6),
    C = c("ctrl", "t0", 7, 8, 9),
    D = c("ctrl", "t1", 10, 11, 12),
    check.names = FALSE
)

# Example 1: extract numeric block using Excel letters, build headers from
# the two rows above (they get collapsed like "cond_t0", "ctrl_t1", ...)
m1 <- extract_data(
    data = df,
    top_row = 3,
    bottom_row = 5,
    left_col = "A",
    right_col = "D",
    feature_name_columns = c("feat_id", "feat_sym"),
    use_row_index = FALSE
)
#> Warning: NAs introduced by coercion
#> Warning: NAs introduced by coercion
m1
#>         X1 X2 cond_t0 cond_t1
#> g1_TP53 NA NA       1       4
#> g2_EGFR NA NA       2       5
#> g3_BAX  NA NA       3       6

# Example 2: same extraction but with numeric column indices and row index
# prepended to ensure uniqueness of feature names
m2 <- extract_data(
    data = df,
    top_row = 3,
    bottom_row = 5,
    left_col = 3,
    right_col = 6,
    feature_name_columns = c("feat_id", "feat_sym"),
    use_row_index = TRUE
)
m2
#>           cond_t0 cond_t1 ctrl_t0 ctrl_t1
#> 1_g1_TP53       1       4       7      10
#> 2_g2_EGFR       2       5       8      11
#> 3_g3_BAX        3       6       9      12