Source (original):<\/strong> DART (Data Analysis, Retrieval and Transfer System), run by Korea’s Financial Supervisory Service \u2014 the mandatory disclosure filing system, same role as SEC EDGAR. Free OpenAPI, 20,000 calls\/day on a free key, no commercial-use restriction.<\/p>\n What’s in it.<\/strong> Every KOSPI\/KOSDAQ listed company’s financial statement line items from annual reports: balance sheet, income statement, cash flow, equity changes. Each row carries the current year amount plus the prior two years side by side, so you get a 3-year trend per line without extra calls. Consolidated (CFS) preferred, falls back to separate (OFS) for companies with no subsidiaries.<\/p>\n The part worth posting about:<\/strong> DART already tags every line with an IFRS taxonomy id \u2014 Gotchas<\/strong> (cost me actual debugging time):<\/p>\n Real example \u2014 Samsung Electronics (005930), FY2022\u20132024, annual consolidated:<\/strong><\/p>\n That FY2023 revenue dip and FY2024 recovery is the memory-chip downturn\/rebound showing up directly in the raw filing data, three years in one API response, no crawling three separate annual reports.<\/p>\n Disclosure \u2014 I built a free wrapper.<\/strong> Raw DART responses are one financial-statement\/report\/year combo per call, Korean-first field names ( The source of truth is DART, not me \u2014 if you’d rather hit opendart.fss.or.kr yourself, the gotcha list above should save you the debugging session it cost me.<\/p>\n<\/div>\n submitted by \/u\/porche911_kr <\/a> \n
ifrs-full_Assets<\/code>, ifrs-full_Revenue<\/code>, ifrs-full_ProfitLoss<\/code>, etc. \u2014 the same taxonomy US\/EU filers use. That means no Korean-to-English account-name translation step<\/strong>, unlike most Korea data sources (my other actor, the real-estate one, needs a full translation layer). You get the Korean label (account_nm<\/code>, e.g. \ub9e4\ucd9c\uc561<\/code>) and the international taxonomy id side by side for free.<\/p>\n\n
corpCode.xml<\/code> (a zip, ~forgot-to-check-Content-Type gotcha), which maps every DART-registered entity (listed and<\/em> unlisted) to an internal corp_code<\/code>. Unlisted entities have an empty stock_code<\/code> field; filter those out or you’ll query companies with no ticker.<\/li>\nfs_div=CFS<\/code> first and fall back to OFS<\/code> \u2014 a normal, expected branch, not an error path. A company with no subsidiaries simply has no CFS.<\/li>\n013<\/code> (“\uc870\ud68c\ub41c \ub370\uc774\ud0c0\uac00 \uc5c6\uc2b5\ub2c8\ub2e4” \/ no data found) is not an error, it’s an empty result \u2014 some report\/year combos genuinely don’t exist. Status 020<\/code> (quota exceeded) is<\/em> a real error. Conflating the two silently turns a rate-limit hit into “this company has no 2019 filings.”<\/li>\n\"514531948000000\"<\/code>), not numbers \u2014 cast, don’t trust the JSON type.<\/li>\nrcept_no<\/code> (the filing receipt number) encodes the filing date in its first 8 digits \u2014 that’s the only place you get an as-of date; there’s no separate date field on the row.<\/li>\n<\/ol>\n\n\n
\n \nFY2022 (\uc81c54\uae30)<\/th>\n FY2023 (\uc81c55\uae30)<\/th>\n FY2024 (\uc81c56\uae30)<\/th>\n<\/tr>\n<\/thead>\n \n Total assets<\/td>\n 448.4T KRW<\/td>\n 455.9T KRW<\/td>\n<\/tr>\n \n Revenue<\/td>\n 302.2T KRW<\/td>\n 258.9T KRW<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n thstrm_amount<\/code>, sj_div<\/code>), and require the corp_code lookup above. I published an Apify actor that does the corp_code mapping, CFS\u2192OFS fallback, and reshapes rows into English field names (currentAmountKrw<\/code>, statementType<\/code>, accountId<\/code>) \u2014 free, no API key of your own needed:<\/p>\n\n
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
