{"id":6322,"date":"2025-02-12T17:18:48","date_gmt":"2025-02-12T08:18:48","guid":{"rendered":"https:\/\/blog.since2020.jp\/?p=6322"},"modified":"2025-02-12T17:19:10","modified_gmt":"2025-02-12T08:19:10","slug":"dataiku-python-time","status":"publish","type":"post","link":"https:\/\/since2020.jp\/media\/dataiku-python-time\/","title":{"rendered":"Dataiku\u3067\u306e\u6642\u7cfb\u5217\u4e88\u6e2c\u306b\u304a\u3051\u308bPython\u51e6\u7406\u306e\u4f8b"},"content":{"rendered":"\n

\u672c\u8a18\u4e8b\u3067\u306f\u3001Dataiku\u3092\u7528\u3044\u305f\u6642\u7cfb\u5217\u4e88\u6e2c\u306e\u30d7\u30ed\u30bb\u30b9\u306b\u304a\u3044\u3066\u3001Python\u6a5f\u80fd\u3092\u6d3b\u7528\u3057\u3066\u30c7\u30fc\u30bf\u306e\u6e96\u5099\u304a\u3088\u3073\u5206\u5272\u3092\u884c\u3046\u65b9\u6cd5\u3092\u89e3\u8aac\u3057\u307e\u3059\u3002\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u51e6\u7406\u3057\u3001\u52b9\u679c\u7684\u306a\u4e88\u6e2c\u3092\u884c\u3046\u305f\u3081\u306b\u306f\u3001\u5b66\u7fd2\u30c7\u30fc\u30bf\u3068\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u5206\u5272\u3059\u308b\u3053\u3068\u304c\u91cd\u8981\u3067\u3059\u3002\r\n<\/p>\n\n\n

\u306f\u3058\u3081\u306b<\/h2>\n

\u672c\u8a18\u4e8b\u3067\u306f\u3001Dataiku\u3092\u7528\u3044\u305f\u6642\u7cfb\u5217\u4e88\u6e2c\u306e\u30d7\u30ed\u30bb\u30b9\u306b\u304a\u3044\u3066\u3001Python\u6a5f\u80fd\u3092\u6d3b\u7528\u3057\u3066\u30c7\u30fc\u30bf\u306e\u6e96\u5099\u304a\u3088\u3073\u5206\u5272\u3092\u884c\u3046\u65b9\u6cd5\u3092\u89e3\u8aac\u3057\u307e\u3059\u3002\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u51e6\u7406\u3057\u3001\u52b9\u679c\u7684\u306a\u4e88\u6e2c\u3092\u884c\u3046\u305f\u3081\u306b\u306f\u3001\u5b66\u7fd2\u30c7\u30fc\u30bf\u3068\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u5206\u5272\u3059\u308b\u3053\u3068\u304c\u91cd\u8981\u3067\u3059\u3002<\/span><\/p>\n\n

1. \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306e\u751f\u6210<\/h2>\n

\u6642\u7cfb\u5217\u4e88\u6e2c\u3067\u306f\u3001\u5c06\u6765\u306e\u30c7\u30fc\u30bf\u3092\u4e8b\u524d\u306b\u751f\u6210\u3057\u3001\u30e2\u30c7\u30eb\u304c\u3069\u306e\u3088\u3046\u306a\u30d1\u30bf\u30fc\u30f3\u3092\u4e88\u6e2c\u3059\u308b\u304b\u3092\u691c\u8a3c\u3059\u308b\u3053\u3068\u304c\u91cd\u8981\u3067\u3059\u3002Code recipes\u3092\u958b\u304dpython\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3059\u3002<\/span><\/p>\r\n

\"\"<\/span><\/p>\r\n

Input\u306b\u52a0\u5de5\u3057\u305f\u3044\u30c7\u30fc\u30bf\u3001Output\u306e\u52a0\u5de5\u5f8c\u306b\u51fa\u529b\u3059\u308b\u305f\u3081\u306e\u30c6\u30fc\u30d6\u30eb\u3092\u7528\u610f\u3057\u307e\u3059\u3002<\/span><\/p>\r\n

\"\"<\/span><\/p>\r\n

\u305d\u306e\u5f8cPyhon\u306e\u30a8\u30c7\u30a3\u30bf\u304c\u5c55\u958b\u3055\u308c\u3001\u30b3\u30fc\u30c9\u3092\u5165\u529b\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u4ee5\u4e0b\u306e\u30b3\u30fc\u30c9\u3067\u306f\u3001\u65e2\u5b58\u306e\u30c7\u30fc\u30bf\u3092\u57fa\u306b\u672a\u6765\u306e7\u65e5\u9593\u306e\u30c7\u30fc\u30bf\u3092\u751f\u6210\u3057\u307e\u3059\u3002\u30b3\u30fc\u30c9\u306e\u5165\u529b\u304c\u5b8c\u4e86\u3057\u305f\u3089Run\u30dc\u30bf\u30f3\u3092\u6559\u3048\u3066\u3001\u51e6\u7406\u3092\u958b\u59cb\u3057\u307e\u3059\u3002<\/span><\/p>\r\n

\r\n
import dataiku\r\nimport pandas as pd, numpy as np\r\nfrom dataiku import pandasutils as pdu\r\n\r\n# Read recipe inputs\r\nhistorical_data = dataiku.Dataset(\"historical_data\")\r\nhistorical_data_df = historical_data.get_dataframe()\r\n\r\n# 'data' \u5217\u3092\u65e5\u4ed8\u578b\u306b\u5909\u63db\r\nhistorical_data_df['data'] = pd.to_datetime(historical_data_df['data'])\r\n\r\n# 'data' \u3067\u30c7\u30fc\u30bf\u3092\u6607\u9806\u306b\u30bd\u30fc\u30c8\r\nhistorical_data_df = historical_data_df.sort_values(by='data')\r\n\r\n# \u6700\u65b0\u306e\u65e5\u4ed8\u3092\u53d6\u5f97\r\nlatest_date = historical_data_df['data'].max()\r\n\r\n# \u672a\u6765\u306e\u30c7\u30fc\u30bf\u65e5\u4ed8\u7bc4\u56f2\u3092\u751f\u6210\uff08\u6700\u65b0\u65e5\u4ed8\u304b\u3089\u6b21\u306e\u65e5\u4ed8\u4ee5\u964d\u306e7\u65e5\u5206\uff09\r\ntest_data_dates = pd.date_range(start=latest_date + pd.Timedelta(days=1), periods=7)\r\n\r\n# \u672a\u6765\u30c7\u30fc\u30bf\u7528\u306b\u30ab\u30c6\u30b4\u30ea\u3092\u53d6\u5f97\r\nunique_categories = historical_data_df['category'].unique()\r\n\r\n# \u672a\u6765\u306e\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u3092\u4f5c\u6210\r\ntest_data = pd.DataFrame([(date, category) for date in test_data_dates for category in unique_categories],\r\ncolumns=['data', 'category'])\r\n\r\n# \u672a\u6765\u30c7\u30fc\u30bf\u306b\u5143\u306e\u30c7\u30fc\u30bf\u306e\u30ab\u30e9\u30e0\u3092\u8ffd\u52a0\u3057\u3001\u5024\u3092\u7a7a\u306b\u3059\u308b\r\ntest_data = test_data.reindex(columns=historical_data_df.columns, fill_value=None)\r\n\r\n# \u5143\u306e\u30c7\u30fc\u30bf\u3068\u672a\u6765\u30c7\u30fc\u30bf\u3092\u7d50\u5408\r\ncombined_data = pd.concat([historical_data_df, test_data], ignore_index=True)\r\n\r\n# \u51fa\u529b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u4ee3\u5165\r\ngenerated_data_df = combined_data\r\n\r\n# Write recipe outputs\r\ngenerated_data = dataiku.Dataset(\"generated_data\")\r\ngenerated_data.write_with_schema(generated_data_df)\r\n<\/code><\/pre>\r\n<\/div>\r\n
\u3053\u306e\u30b3\u30fc\u30c9\u3067\u306f\u3001\u4ee5\u4e0b\u306e\u51e6\u7406\u3092\u884c\u3063\u3066\u3044\u307e\u3059\u3002<\/span><\/p>\r\n
    \r\n\t
  • \u30c7\u30fc\u30bf\u306e\u8aad\u307f\u8fbc\u307f<\/strong>\uff1aDataiku\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u53d6\u5f97<\/span><\/li>\r\n\t
  • \u65e5\u4ed8\u306e\u5909\u63db\u3068\u30bd\u30fc\u30c8<\/strong>\uff1a\u30c7\u30fc\u30bf\u3092\u6642\u7cfb\u5217\u9806\u306b\u4e26\u3079\u308b<\/span><\/li>\r\n\t
  • \u672a\u6765\u306e\u30c7\u30fc\u30bf\u751f\u6210<\/strong>\uff1a\u6700\u65b0\u65e5\u4ed8\u304b\u30897\u65e5\u5206\u306e\u672a\u6765\u30c7\u30fc\u30bf\u3092\u4f5c\u6210<\/span><\/li>\r\n\t
  • \u30ab\u30c6\u30b4\u30ea\u306e\u9069\u7528<\/strong>\uff1a\u5143\u306e\u30c7\u30fc\u30bf\u306e\u30ab\u30c6\u30b4\u30ea\u3092\u672a\u6765\u30c7\u30fc\u30bf\u306b\u3082\u9069\u7528<\/span><\/li>\r\n\t
  • \u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306e\u4fdd\u5b58<\/strong>\uff1a\u751f\u6210\u3057\u305f\u30c7\u30fc\u30bf\u3092Dataiku\u306b\u66f8\u304d\u8fbc\u3080<\/span><\/li>\r\n<\/ul>\r\n
    \u3053\u306e\u51e6\u7406\u306b\u3088\u308a\u3001\u4e88\u6e2c\u30e2\u30c7\u30eb\u304c\u5b66\u7fd2\u3059\u308b\u305f\u3081\u306e\u9069\u5207\u306a\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u3092\u6e96\u5099\u3067\u304d\u307e\u3059\u3002<\/span><\/p>\r\n
    <\/span><\/p>\r\n
    \"\"<\/p>\r\n
    <\/p>\n\n
    2. \u30c7\u30fc\u30bf\u306e\u5206\u5272<\/h2>\n
    \u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u5b66\u7fd2\u30fb\u691c\u8a3c\u30fb\u30c6\u30b9\u30c8\u306e3\u3064\u306e\u671f\u9593\u306b\u5206\u5272\u3059\u308b\u3053\u3068\u3067\u3001\u30e2\u30c7\u30eb\u306e\u8a55\u4fa1\u7cbe\u5ea6\u3092\u5411\u4e0a\u3055\u305b\u307e\u3059\u30021\u3068\u540c\u3058\u3088\u3046\u306bPython\u3092\u958b\u304d\u3001\u30b3\u30fc\u30c9\u3092\u5b9f\u884c\u3057\u3066\u3044\u304d\u307e\u3059\u3002\u4ee5\u4e0b\u306e\u30b3\u30fc\u30c9\u3067\u306f\u3001\u5b66\u7fd2\u671f\u9593\u3092365\u65e5\uff081\u5e74\uff09\u3001\u691c\u8a3c\u671f\u9593\u30927\u65e5\u3001\u30c6\u30b9\u30c8\u671f\u9593\u30927\u65e5\u3068\u8a2d\u5b9a\u3057\u3001\u30c7\u30fc\u30bf\u3092\u5206\u5272\u3057\u307e\u3059\u3002<\/span><\/p>\r\n
    \r\n
    import dataiku\r\nimport pandas as pd\r\nfrom dataiku import pandasutils as pdu\r\n\r\n# Dataiku\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u8aad\u307f\u8fbc\u3080\r\nprepared_data = dataiku.Dataset(\"prepared_data\")\r\ndf = prepared_data.get_dataframe()\r\n\r\n# 'data' \u3092\u65e5\u4ed8\u578b\u306b\u5909\u63db\r\ndf['data'] = pd.to_datetime(df['data'])\r\n\r\n# \u5e74\u672b\u5e74\u59cb\u306e\u671f\u9593\uff0812\u670823\u65e5\uff5e1\u67085\u65e5\uff09\u3092\u9664\u5916\r\ndf = df[~(\r\n((df['data'].dt.month == 12) & (df['data'].dt.day >= 23)) |\r\n((df['data'].dt.month == 1) & (df['data'].dt.day <= 5))\r\n)]\r\n\r\n# 'data' \u3067\u30c7\u30fc\u30bf\u3092\u6607\u9806\u306b\u30bd\u30fc\u30c8\r\ndf = df.sort_values(by='data')\r\n\r\n# \u5404\u671f\u9593\u306e\u65e5\u6570\u3092\u8a2d\u5b9a\r\ntest_period_days = 7 # \u30c6\u30b9\u30c8\uff1a\u6700\u65b01\u9031\u9593\r\nvalidate_period_days = 7 # \u691c\u8a3c\uff1a\u30c6\u30b9\u30c8\u76f4\u524d\u306e1\u9031\u9593\r\ntrain_period_days = 365 # \u5b66\u7fd2\uff1a\u691c\u8a3c\u76f4\u524d\u306e1\u5e74\u9593\r\n\r\n# \u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u5185\u306e\u6700\u65b0\u306e\u65e5\u4ed8\u3092\u53d6\u5f97\r\nlatest_date = df['data'].max()\r\n\r\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306e\u62bd\u51fa\r\ntest_data_start_date = latest_date - pd.Timedelta(days=test_period_days - 1)\r\ntest_data_end_date = latest_date\r\ntest_data = df[\r\n(df['data'] >= test_data_start_date) &\r\n(df['data'] <= test_data_end_date)\r\n]\r\n\r\n# \u691c\u8a3c\u30c7\u30fc\u30bf\u306e\u62bd\u51fa\r\nvalidate_data_end_date = test_data_start_date - pd.Timedelta(days=1)\r\nvalidate_data_start_date = validate_data_end_date - pd.Timedelta(days=validate_period_days - 1)\r\nvalidate_data = df[\r\n(df['data'] >= validate_data_start_date) &\r\n(df['data'] <= validate_data_end_date)\r\n]\r\n\r\n# \u5b66\u7fd2\u30c7\u30fc\u30bf\u306e\u62bd\u51fa\r\ntrain_data_end_date = validate_data_start_date - pd.Timedelta(days=1)\r\ntrain_data_start_date = train_data_end_date - pd.Timedelta(days=train_period_days - 1)\r\ntrain_data = df[\r\n(df['data'] >= train_data_start_date) &\r\n(df['data'] <= train_data_end_date)\r\n]\r\n\r\n# Dataiku\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3078\u306e\u66f8\u304d\u8fbc\u307f\r\ntrain_dataset = dataiku.Dataset(\"train_dataset\")\r\ntrain_dataset.write_with_schema(train_data)\r\n\r\nvalidate_dataset = dataiku.Dataset(\"validate_dataset\")\r\nvalidate_dataset.write_with_schema(validate_data)\r\n\r\ntest_dataset = dataiku.Dataset(\"test_dataset\")\r\ntest_dataset.write_with_schema(test_data)\r\n<\/code><\/pre>\r\n<\/div>\r\n
    \u3053\u306e\u30b3\u30fc\u30c9\u3067\u306f\u3001<\/span><\/p>\r\n
    1.\u30c7\u30fc\u30bf\u306e\u53d6\u5f97\u3068\u524d\u51e6\u7406<\/strong><\/span><\/p>\r\n
    2.\u5b66\u7fd2\u30fb\u691c\u8a3c\u30fb\u30c6\u30b9\u30c8\u306e3\u3064\u306e\u671f\u9593\u306b\u5206\u5272<\/strong><\/span><\/p>\r\n
    3.Dataiku\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3068\u3057\u3066\u4fdd\u5b58<\/strong> \u3092\u5b9f\u884c\u3057\u3066\u3044\u307e\u3059\u3002<\/span><\/p>\r\n
    <\/p>\r\n
    <\/p>\n\n
    \u307e\u3068\u3081<\/h2>\n
    \u672c\u8a18\u4e8b\u3067\u306f\u3001Dataiku\u3067\u306e\u6642\u7cfb\u5217\u4e88\u6e2c\u306b\u304a\u3051\u308bPython\u3092\u6d3b\u7528\u3057\u305f\u30c7\u30fc\u30bf\u51e6\u7406\u306b\u3064\u3044\u3066\u89e3\u8aac\u3057\u307e\u3057\u305f\u3002\u5b66\u7fd2\u671f\u95931\u5e74\u3001\u4e88\u6e2c\u671f\u95931\u9031\u9593\u3068\u3044\u3046\u8a2d\u5b9a\u306b\u3088\u308a\u3001\u77ed\u671f\u4e88\u6e2c\u306b\u9069\u3057\u305f\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u4f5c\u6210\u3055\u308c\u4eca\u5f8c\u306e\u904b\u7528\u306b\u9069\u3057\u305f\u30d5\u30ed\u30fc\u3068\u306a\u308a\u307e\u3057\u305f\u3002Dataiku\u306e\u30d3\u30b8\u30e5\u30a2\u30eb\u6a5f\u80fd\u3068Python\u3092\u7d44\u307f\u5408\u308f\u305b\u308b\u3053\u3068\u3067\u3001\u3088\u308a\u67d4\u8edf\u306a\u30c7\u30fc\u30bf\u51e6\u7406\u3068\u4e88\u6e2c\u30e2\u30c7\u30eb\u306e\u958b\u767a\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002<\/span><\/p>\r\n
    <\/p>","protected":false},"excerpt":{"rendered":"
    \u672c\u8a18\u4e8b\u3067\u306f\u3001Dataiku\u3092\u7528\u3044\u305f\u6642\u7cfb\u5217\u4e88\u6e2c\u306e\u30d7\u30ed\u30bb\u30b9\u306b\u304a\u3044\u3066\u3001Python\u6a5f\u80fd\u3092\u6d3b\u7528\u3057\u3066\u30c7\u30fc\u30bf\u306e\u6e96\u5099\u304a\u3088\u3073\u5206\u5272\u3092\u884c\u3046\u65b9\u6cd5\u3092\u89e3\u8aac\u3057\u307e\u3059\u3002\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u3092\u9069\u5207\u306b\u51e6\u7406\u3057\u3001\u52b9\u679c\u7684\u306a\u4e88\u6e2c\u3092\u884c\u3046\u305f\u3081\u306b\u306f\u3001\u5b66\u7fd2\u30c7\u30fc\u30bf\u3068\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u3092\u9069 […]<\/p>\n","protected":false},"author":87,"featured_media":3371,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","swell_btn_cv_data":"","footnotes":"","_wp_rev_ctl_limit":""},"categories":[1246],"tags":[199,331,552],"class_list":["post-6322","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-data-infrastructure","tag-dataiku","tag-python","tag-552"],"_links":{"self":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts\/6322","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/users\/87"}],"replies":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/comments?post=6322"}],"version-history":[{"count":0,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts\/6322\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/media\/3371"}],"wp:attachment":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/media?parent=6322"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/categories?post=6322"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/tags?post=6322"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}