Reading and Writing Files

File access in Python is managed through a file object obtained from open. The object holds an operating-system resource that must be released; leaving it open risks exhausting the process's file handles and, for files opened for writing, losing buffered data that was never flushed.

The with statement guarantees release. It closes the file when the block ends, including when the block exits because of an exception. Opening a file outside a with block, without a matching close, is a defect.

gene	control_1	control_2	treated_1	treated_2

Several files are already mounted in this sandbox, listed in the Files tab. counts.tsv and variants.tsv are used in the examples below.

Reading line by line

Iterating over a file object yields one line at a time and reads only as much as required, so it works on files far larger than available memory. This is the standard approach for large datasets, as high-throughput sequencing output routinely exceeds available physical RAM.

Each line retains its trailing newline, so strip is almost always applied before use.

chrom	pos	ref	alt	gene	qual
chr17	7676154	G	A	TP53	99.0
chr17	43093220	C	T	BRCA1	45.2
chr7	55191822	T	G	EGFR	88.7
chr12	25245350	C	T	KRAS	12.4
chr7	140753336	A	T	BRAF	76.1

Reading the whole file

read returns the entire contents as one string, and readlines returns a list of lines. Both load the whole file into memory and are appropriate only when the size is known to be small.

7 lines
gene	control_1	control_2	treated_1	treated_2

splitlines is preferred over split("\n") because it removes the line terminators and handles the carriage-return conventions of files produced on other platforms.

Parsing delimited data

A tab-separated line becomes a list of fields with split("\t"). Read the header row first, then use its index positions to locate columns by name rather than by a hard-coded number, so that the code survives a change of column order.

TP53 2310
BRCA1 441
EGFR 1502
KRAS 1760
MYC 900
GAPDH 15100

Writing

The second argument to open is the mode. "r" reads and is the default; "w" creates the file or truncates an existing one; "a" appends. "w" discards existing contents without warning.

gene	depth
TP53	120

write does not append a newline; it must be supplied explicitly. Note that the sandbox filesystem is temporary and is discarded when the page reloads.

Task

Write column_sum(path, column). Read a tab-separated file with a header row and return the sum of the named column as an integer. Assume the column exists and every value in it is a whole number.

Locate the column with header.index(column) rather than assuming a position, and skip blank lines before splitting, since splitting an empty line produces a single empty field that will not convert to an integer.