Stata Log File Parser — Free Command & Output Extractor

This tool parses a Stata log file pasted as text and extracts the commands that were run, in order, with a summary of how many command lines and output lines the log contains. Stata logs interleave commands (lines beginning with “. ”) and their output, so reconstructing what an analysis actually did means reading through pages of tables. The parser separates the two: the command list becomes a readable record of the analysis — the do-file that the log implies — while output lines are counted but not reproduced. That command list is the reproducibility artifact: it shows the steps, in order, without the noise.

How to use it

Paste the Stata log text.
Click Parse log.
Read the command list in order.
Compare it with the do-file you meant to run.
Save the command list as the analysis record.

Worked example

A log contains . use data.dta, . summarize age (with its output table), and . regress y x. The parser returns three commands in order — use data.dta, summarize age, regress y x — and counts the output lines separately. A student checking a classmate’s analysis can now see in seconds that the regression ran after the summary and that no data-cleaning command preceded it, which raises the right question: were missing values or outliers handled before modelling, or was the raw file regressed directly? The log answers what happened; the command list makes the answer readable.

Honest limits

The parser recognizes commands only by the “. ” prefix in pasted text; logs with different formats (SMCL markup, logs without the command echo, or commands split across continuation lines) may parse incompletely, and it does not capture command options that span lines or the contents of output tables. It also cannot verify that the commands produced the claimed results — a log can be edited, and a command list is only as trustworthy as the log it came from. For a formal audit trail, keep the original log file and the do-file together, and treat the parsed list as a reading aid, not as a replacement for either.

Who it is for

This tool is for Stata users reviewing logs, students learning what an analysis actually consisted of, and teams reconstructing a do-file from a log when the original script is lost. It turns a wall of output into a story of steps.

Going further

Make logs useful by construction: run analyses from do-files rather than interactively, log every run, and never edit a log by hand — an edited log is worse than no log, because it claims an authority it no longer has. Name log files with the analysis and date, and archive them with the data version they ran on, so a future reader can match log, code, and data without detective work. When a log and a do-file disagree, believe neither until you rerun: the disagreement itself is the finding, and resolving it — a changed file, a forgotten step — is often where the real analysis error hides. Compare the parsed command list against the analysis plan, not just against the do-file: a command that ran but was never planned, or a planned step that never ran, reveals the gap between intention and execution that matters most for reproducibility. Keep the comparison as a short note with the log — planned steps, executed steps, and explanations for any difference — because a future reader who finds the note understands the analysis, while one who finds only the log must guess what was intended. Intention plus execution is the complete record. Archive the log with its data and code as a trio, and name them consistently — same project code, same date stamp — so any one of the three leads to the other two. A log without its data proves nothing about results, and data without its log proves nothing about process; together with the do-file they form a complete reproducibility package that a reviewer can actually rerun. That trio, consistently named and stored together, is the standard every analysis should meet before it is called finished.

Embed this tool on your website — free

Copy the code below to add this working tool to your own page. A small credit link stays with it — that’s all we ask.