質問編集履歴
2
クエリ例の追加
test
CHANGED
|
File without changes
|
test
CHANGED
|
@@ -165,3 +165,163 @@
|
|
|
165
165
|
SQLに詳しい方、ビックデータに詳しい方etc...
|
|
166
166
|
|
|
167
167
|
ご教授いただけますと幸いですb
|
|
168
|
+
|
|
169
|
+
|
|
170
|
+
|
|
171
|
+
|
|
172
|
+
|
|
173
|
+
# 20160728追記
|
|
174
|
+
|
|
175
|
+
* 一応、window関数とWHERE句を使って書くことはできました(以下になります)。
|
|
176
|
+
|
|
177
|
+
* まだ冗長なクエリなので、何かスマートな書き方をご存知な方がいれば、教えていただけますと幸いです...b
|
|
178
|
+
|
|
179
|
+
|
|
180
|
+
|
|
181
|
+
```sql
|
|
182
|
+
|
|
183
|
+
-- 日別プラットフォーム別に、課金したユーザの一覧を取る
|
|
184
|
+
|
|
185
|
+
WITH sub1 AS
|
|
186
|
+
|
|
187
|
+
(
|
|
188
|
+
|
|
189
|
+
SELECT
|
|
190
|
+
|
|
191
|
+
TD_TIME_FORMAT(time,'yyyy-MM-dd','JST') AS date
|
|
192
|
+
|
|
193
|
+
,{platform}
|
|
194
|
+
|
|
195
|
+
,user_id
|
|
196
|
+
|
|
197
|
+
FROM
|
|
198
|
+
|
|
199
|
+
kpi_table
|
|
200
|
+
|
|
201
|
+
WHERE
|
|
202
|
+
|
|
203
|
+
--月初め〜月末まで取る
|
|
204
|
+
|
|
205
|
+
TD_TIME_RANGE(
|
|
206
|
+
|
|
207
|
+
time
|
|
208
|
+
|
|
209
|
+
,TD_TIME_FORMAT({month_start},'yyyy-MM-dd 00:00:00','JST')
|
|
210
|
+
|
|
211
|
+
,TD_TIME_FORMAT({month_last},'yyyy-MM-dd 23:59:59','JST')
|
|
212
|
+
|
|
213
|
+
,'JST'
|
|
214
|
+
|
|
215
|
+
)
|
|
216
|
+
|
|
217
|
+
--課金したユーザのみ
|
|
218
|
+
|
|
219
|
+
AND (purchase_amount > 0)
|
|
220
|
+
|
|
221
|
+
GROUP BY
|
|
222
|
+
|
|
223
|
+
TD_TIME_FORMAT(time,'yyyy-MM-dd','JST')
|
|
224
|
+
|
|
225
|
+
,{platform}
|
|
226
|
+
|
|
227
|
+
,user_id
|
|
228
|
+
|
|
229
|
+
ORDER BY
|
|
230
|
+
|
|
231
|
+
TD_TIME_FORMAT(time,'yyyy-MM-dd','JST') ASC
|
|
232
|
+
|
|
233
|
+
,{platform} ASC
|
|
234
|
+
|
|
235
|
+
,user_id
|
|
236
|
+
|
|
237
|
+
)
|
|
238
|
+
|
|
239
|
+
-- ユーザ毎のログイン日を集計
|
|
240
|
+
|
|
241
|
+
, sub2 AS
|
|
242
|
+
|
|
243
|
+
(
|
|
244
|
+
|
|
245
|
+
SELECT
|
|
246
|
+
|
|
247
|
+
date
|
|
248
|
+
|
|
249
|
+
,{platform}
|
|
250
|
+
|
|
251
|
+
,user_id
|
|
252
|
+
|
|
253
|
+
,COUNT(user_id) OVER (PARTITION BY {platform}, user_id ORDER BY date) AS login_count
|
|
254
|
+
|
|
255
|
+
FROM
|
|
256
|
+
|
|
257
|
+
sub1
|
|
258
|
+
|
|
259
|
+
ORDER BY
|
|
260
|
+
|
|
261
|
+
date ASC
|
|
262
|
+
|
|
263
|
+
,{platform} ASC
|
|
264
|
+
|
|
265
|
+
,user_id
|
|
266
|
+
|
|
267
|
+
)
|
|
268
|
+
|
|
269
|
+
-- 全ユーザのログイン日を集計
|
|
270
|
+
|
|
271
|
+
, sub3 AS
|
|
272
|
+
|
|
273
|
+
(
|
|
274
|
+
|
|
275
|
+
SELECT
|
|
276
|
+
|
|
277
|
+
date
|
|
278
|
+
|
|
279
|
+
,{platform}
|
|
280
|
+
|
|
281
|
+
,COUNT(user_id) OVER (PARTITION BY platform ORDER BY date) user_num
|
|
282
|
+
|
|
283
|
+
FROM
|
|
284
|
+
|
|
285
|
+
sub2
|
|
286
|
+
|
|
287
|
+
--最初にログインした日以外のデータは排除
|
|
288
|
+
|
|
289
|
+
WHERE
|
|
290
|
+
|
|
291
|
+
login_count = 1
|
|
292
|
+
|
|
293
|
+
ORDER BY
|
|
294
|
+
|
|
295
|
+
date ASC
|
|
296
|
+
|
|
297
|
+
,{platform} ASC
|
|
298
|
+
|
|
299
|
+
)
|
|
300
|
+
|
|
301
|
+
-- データを日別 * プラットフォーム別に絞る
|
|
302
|
+
|
|
303
|
+
SELECT
|
|
304
|
+
|
|
305
|
+
date
|
|
306
|
+
|
|
307
|
+
,{platform}
|
|
308
|
+
|
|
309
|
+
MAX(user_num) AS user_num
|
|
310
|
+
|
|
311
|
+
FROM
|
|
312
|
+
|
|
313
|
+
sub3
|
|
314
|
+
|
|
315
|
+
GROUP BY
|
|
316
|
+
|
|
317
|
+
date
|
|
318
|
+
|
|
319
|
+
,{platform}
|
|
320
|
+
|
|
321
|
+
ORDER BY
|
|
322
|
+
|
|
323
|
+
date ASC
|
|
324
|
+
|
|
325
|
+
,{platform} ASC
|
|
326
|
+
|
|
327
|
+
```
|
1
GROUP BY内の抜けを修正しました
test
CHANGED
|
File without changes
|
test
CHANGED
|
@@ -92,6 +92,8 @@
|
|
|
92
92
|
|
|
93
93
|
,{platform}
|
|
94
94
|
|
|
95
|
+
,user_id
|
|
96
|
+
|
|
95
97
|
ORDER BY
|
|
96
98
|
|
|
97
99
|
TD_TIME_FORMAT(time,'yyyy-MM-dd','JST') ASC
|