ワンクリックでManusで任意のスキルを実行

table-extraction

スター3

フォーク1

更新日2026年4月8日 05:27

How to accurately extract values from Treasury Bulletin tables. Use when reading tables with many columns, similar row labels, or when you need to verify you have the right cell.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

jwalin-shah

jwalin-shah/officeqa-arena

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

Accurate Table Extraction

Wide tables (10+ columns) cause misreads. To verify you have the right value:

Identify the EXACT row label asked. "National defense" ≠ "National defense and veterans" ≠ "Total national security"
Count columns from the header to confirm which column has your target year/period
For wide tables, extract just the row you need: grep "row label" file.txt Then count pipe-separated values to match the column header.
Cross-check: if the page file has the value AND the full .txt has it, they should match. If not, trust the page file.

Common Table Traps

"Total" rows already include sub-items. Never sum a parent with its children.
Some tables wrap across pages. Check if there is a "Continued" table on the next page.
Footnote markers (1/, 2/, r/, p/) appear inside numbers — strip them before computing.
Values like "(*)" or "-" mean zero or not applicable, not missing data.
Column headers with ">" indicate multi-level headers flattened. Read carefully.

このリポジトリの他の Skills

同じリポジトリ

verify

jwalin-shah/officeqa-arena

Review board — verify the analyst's answer before final submission

2026-05-143

computation-patterns

jwalin-shah/officeqa-arena

Python computation formulas for Treasury data analysis. Use for percent change, CAGR, stdev, kurtosis, VaR, regression, Theil index, and other statistical calculations.

2026-04-083

cpi-reference

jwalin-shah/officeqa-arena

CPI-U annual averages (1929-2024) for inflation adjustment. Use when question mentions inflation, CPI, constant dollars, or real dollars.

2026-04-083

data-files

jwalin-shah/officeqa-arena

How to read Treasury Bulletin data files in /app/resources/. Use when exploring resources or when JSON parsing fails.

2026-04-083

fiscal-calendar

jwalin-shah/officeqa-arena

U.S. fiscal year boundaries and calendar year conversions. Use when question mentions fiscal year, FY, or when you need to determine which months belong to a fiscal year.

2026-04-083

officeqa-checklist

jwalin-shah/officeqa-arena

Step-by-step approach and verification checklist for Treasury Bulletin questions

2026-04-083

name	table-extraction
description	How to accurately extract values from Treasury Bulletin tables. Use when reading tables with many columns, similar row labels, or when you need to verify you have the right cell.

Accurate Table Extraction

Wide tables (10+ columns) cause misreads. To verify you have the right value:

Identify the EXACT row label asked. "National defense" ≠ "National defense and veterans" ≠ "Total national security"
Count columns from the header to confirm which column has your target year/period
For wide tables, extract just the row you need: grep "row label" file.txt Then count pipe-separated values to match the column header.
Cross-check: if the page file has the value AND the full .txt has it, they should match. If not, trust the page file.

Common Table Traps

"Total" rows already include sub-items. Never sum a parent with its children.
Some tables wrap across pages. Check if there is a "Continued" table on the next page.
Footnote markers (1/, 2/, r/, p/) appear inside numbers — strip them before computing.
Values like "(*)" or "-" mean zero or not applicable, not missing data.
Column headers with ">" indicate multi-level headers flattened. Read carefully.