funcs_agg.go 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580
  1. // Copyright 2022-2023 EMQ Technologies Co., Ltd.
  2. //
  3. // Licensed under the Apache License, Version 2.0 (the "License");
  4. // you may not use this file except in compliance with the License.
  5. // You may obtain a copy of the License at
  6. //
  7. // http://www.apache.org/licenses/LICENSE-2.0
  8. //
  9. // Unless required by applicable law or agreed to in writing, software
  10. // distributed under the License is distributed on an "AS IS" BASIS,
  11. // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  12. // See the License for the specific language governing permissions and
  13. // limitations under the License.
  14. package function
  15. import (
  16. "fmt"
  17. "github.com/montanaflynn/stats"
  18. "github.com/lf-edge/ekuiper/pkg/api"
  19. "github.com/lf-edge/ekuiper/pkg/ast"
  20. "github.com/lf-edge/ekuiper/pkg/cast"
  21. )
  22. func registerAggFunc() {
  23. builtins["avg"] = builtinFunc{
  24. fType: ast.FuncTypeAgg,
  25. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  26. arg0 := args[0].([]interface{})
  27. c := getCount(arg0)
  28. if c > 0 {
  29. v := getFirstValidArg(arg0)
  30. switch v.(type) {
  31. case int, int64:
  32. if r, err := sliceIntTotal(arg0); err != nil {
  33. return err, false
  34. } else {
  35. return r / int64(c), true
  36. }
  37. case float64:
  38. if r, err := sliceFloatTotal(arg0); err != nil {
  39. return err, false
  40. } else {
  41. return r / float64(c), true
  42. }
  43. case nil:
  44. return nil, true
  45. default:
  46. return fmt.Errorf("run avg function error: found invalid arg %[1]T(%[1]v)", v), false
  47. }
  48. }
  49. return nil, true
  50. },
  51. val: ValidateOneNumberArg,
  52. check: returnNilIfHasAnyNil,
  53. }
  54. builtins["count"] = builtinFunc{
  55. fType: ast.FuncTypeAgg,
  56. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  57. arg0 := args[0].([]interface{})
  58. return getCount(arg0), true
  59. },
  60. val: ValidateOneArg,
  61. check: returnNilIfHasAnyNil,
  62. }
  63. builtins["max"] = builtinFunc{
  64. fType: ast.FuncTypeAgg,
  65. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  66. arg0 := args[0].([]interface{})
  67. if len(arg0) > 0 {
  68. v := getFirstValidArg(arg0)
  69. switch t := v.(type) {
  70. case int:
  71. if r, err := sliceIntMax(arg0, int64(t)); err != nil {
  72. return err, false
  73. } else {
  74. return r, true
  75. }
  76. case int64:
  77. if r, err := sliceIntMax(arg0, t); err != nil {
  78. return err, false
  79. } else {
  80. return r, true
  81. }
  82. case float64:
  83. if r, err := sliceFloatMax(arg0, t); err != nil {
  84. return err, false
  85. } else {
  86. return r, true
  87. }
  88. case string:
  89. if r, err := sliceStringMax(arg0, t); err != nil {
  90. return err, false
  91. } else {
  92. return r, true
  93. }
  94. case nil:
  95. return nil, true
  96. default:
  97. return fmt.Errorf("run max function error: found invalid arg %[1]T(%[1]v)", v), false
  98. }
  99. }
  100. return nil, true
  101. },
  102. val: ValidateOneNumberArg,
  103. check: returnNilIfHasAnyNil,
  104. }
  105. builtins["min"] = builtinFunc{
  106. fType: ast.FuncTypeAgg,
  107. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  108. arg0 := args[0].([]interface{})
  109. if len(arg0) > 0 {
  110. v := getFirstValidArg(arg0)
  111. switch t := v.(type) {
  112. case int:
  113. if r, err := sliceIntMin(arg0, int64(t)); err != nil {
  114. return err, false
  115. } else {
  116. return r, true
  117. }
  118. case int64:
  119. if r, err := sliceIntMin(arg0, t); err != nil {
  120. return err, false
  121. } else {
  122. return r, true
  123. }
  124. case float64:
  125. if r, err := sliceFloatMin(arg0, t); err != nil {
  126. return err, false
  127. } else {
  128. return r, true
  129. }
  130. case string:
  131. if r, err := sliceStringMin(arg0, t); err != nil {
  132. return err, false
  133. } else {
  134. return r, true
  135. }
  136. case nil:
  137. return nil, true
  138. default:
  139. return fmt.Errorf("run min function error: found invalid arg %[1]T(%[1]v)", v), false
  140. }
  141. }
  142. return nil, true
  143. },
  144. val: ValidateOneNumberArg,
  145. check: returnNilIfHasAnyNil,
  146. }
  147. builtins["sum"] = builtinFunc{
  148. fType: ast.FuncTypeAgg,
  149. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  150. arg0 := args[0].([]interface{})
  151. if len(arg0) > 0 {
  152. v := getFirstValidArg(arg0)
  153. switch v.(type) {
  154. case int, int64:
  155. if r, err := sliceIntTotal(arg0); err != nil {
  156. return err, false
  157. } else {
  158. return r, true
  159. }
  160. case float64:
  161. if r, err := sliceFloatTotal(arg0); err != nil {
  162. return err, false
  163. } else {
  164. return r, true
  165. }
  166. case nil:
  167. return nil, true
  168. default:
  169. return fmt.Errorf("run sum function error: found invalid arg %[1]T(%[1]v)", v), false
  170. }
  171. }
  172. return nil, true
  173. },
  174. val: ValidateOneNumberArg,
  175. check: returnNilIfHasAnyNil,
  176. }
  177. builtins["collect"] = builtinFunc{
  178. fType: ast.FuncTypeAgg,
  179. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  180. if len(args) > 0 {
  181. return args[0], true
  182. }
  183. return make([]interface{}, 0), true
  184. },
  185. val: ValidateOneArg,
  186. }
  187. builtins["merge_agg"] = builtinFunc{
  188. fType: ast.FuncTypeAgg,
  189. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  190. data, ok := args[0].([]interface{})
  191. if ok {
  192. result := make(map[string]interface{})
  193. for _, ele := range data {
  194. if m, ok := ele.(map[string]interface{}); ok {
  195. for k, v := range m {
  196. result[k] = v
  197. }
  198. }
  199. }
  200. return result, true
  201. }
  202. return nil, true
  203. },
  204. val: ValidateOneArg,
  205. check: returnNilIfHasAnyNil,
  206. }
  207. builtins["deduplicate"] = builtinFunc{
  208. fType: ast.FuncTypeAgg,
  209. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  210. v1, ok1 := args[0].([]interface{})
  211. v2, ok2 := args[1].([]interface{})
  212. v3a, ok3 := args[2].([]interface{})
  213. if ok1 && ok2 && ok3 && len(v3a) > 0 {
  214. v3, ok4 := getFirstValidArg(v3a).(bool)
  215. if ok4 {
  216. if r, err := dedup(v1, v2, v3); err != nil {
  217. return err, false
  218. } else {
  219. return r, true
  220. }
  221. }
  222. }
  223. return fmt.Errorf("Invalid argument type found."), false
  224. },
  225. val: func(_ api.FunctionContext, args []ast.Expr) error {
  226. if err := ValidateLen(2, len(args)); err != nil {
  227. return err
  228. }
  229. if !ast.IsBooleanArg(args[1]) {
  230. return ProduceErrInfo(1, "bool")
  231. }
  232. return nil
  233. },
  234. check: returnNilIfHasAnyNil,
  235. }
  236. builtins["stddev"] = builtinFunc{
  237. fType: ast.FuncTypeAgg,
  238. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  239. arg0 := args[0].([]interface{})
  240. if len(arg0) > 0 {
  241. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  242. if err != nil {
  243. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  244. }
  245. deviation, err := stats.StandardDeviation(float64Slice)
  246. if err != nil {
  247. if err == stats.EmptyInputErr {
  248. return nil, true
  249. }
  250. return fmt.Errorf("StandardDeviation exec with error: %v", err), false
  251. }
  252. return deviation, true
  253. }
  254. return nil, true
  255. },
  256. val: ValidateOneNumberArg,
  257. check: returnNilIfHasAnyNil,
  258. }
  259. builtins["stddevs"] = builtinFunc{
  260. fType: ast.FuncTypeAgg,
  261. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  262. arg0 := args[0].([]interface{})
  263. if len(arg0) > 0 {
  264. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  265. if err != nil {
  266. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  267. }
  268. deviation, err := stats.StandardDeviationSample(float64Slice)
  269. if err != nil {
  270. if err == stats.EmptyInputErr {
  271. return nil, true
  272. }
  273. return fmt.Errorf("StandardDeviationSample exec with error: %v", err), false
  274. }
  275. return deviation, true
  276. }
  277. return nil, true
  278. },
  279. val: ValidateOneNumberArg,
  280. check: returnNilIfHasAnyNil,
  281. }
  282. builtins["var"] = builtinFunc{
  283. fType: ast.FuncTypeAgg,
  284. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  285. arg0 := args[0].([]interface{})
  286. if len(arg0) > 0 {
  287. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  288. if err != nil {
  289. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  290. }
  291. deviation, err := stats.Variance(float64Slice)
  292. if err != nil {
  293. if err == stats.EmptyInputErr {
  294. return nil, true
  295. }
  296. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  297. }
  298. return deviation, true
  299. }
  300. return nil, true
  301. },
  302. val: ValidateOneNumberArg,
  303. check: returnNilIfHasAnyNil,
  304. }
  305. builtins["vars"] = builtinFunc{
  306. fType: ast.FuncTypeAgg,
  307. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  308. arg0 := args[0].([]interface{})
  309. if len(arg0) > 0 {
  310. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  311. if err != nil {
  312. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  313. }
  314. deviation, err := stats.SampleVariance(float64Slice)
  315. if err != nil {
  316. if err == stats.EmptyInputErr {
  317. return nil, true
  318. }
  319. return fmt.Errorf("SampleVariance exec with error: %v", err), false
  320. }
  321. return deviation, true
  322. }
  323. return nil, true
  324. },
  325. val: ValidateOneNumberArg,
  326. check: returnNilIfHasAnyNil,
  327. }
  328. builtins["percentile_cont"] = builtinFunc{
  329. fType: ast.FuncTypeAgg,
  330. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  331. if err := ValidateLen(2, len(args)); err != nil {
  332. return err, false
  333. }
  334. var arg1Float64 float64 = 1
  335. arg0 := args[0].([]interface{})
  336. arg1 := args[1].([]interface{})
  337. if len(arg1) > 0 {
  338. v1 := getFirstValidArg(arg1)
  339. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  340. if err != nil {
  341. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  342. }
  343. arg1Float64 = val
  344. }
  345. if len(arg0) > 0 {
  346. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  347. if err != nil {
  348. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  349. }
  350. deviation, err := stats.Percentile(float64Slice, arg1Float64*100)
  351. if err != nil {
  352. if err == stats.EmptyInputErr {
  353. return nil, true
  354. }
  355. return fmt.Errorf("percentile exec with error: %v", err), false
  356. }
  357. return deviation, true
  358. }
  359. return nil, true
  360. },
  361. val: ValidateTwoNumberArg,
  362. check: returnNilIfHasAnyNil,
  363. }
  364. builtins["percentile_disc"] = builtinFunc{
  365. fType: ast.FuncTypeAgg,
  366. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  367. if err := ValidateLen(2, len(args)); err != nil {
  368. return err, false
  369. }
  370. var arg1Float64 float64 = 1
  371. arg0 := args[0].([]interface{})
  372. arg1 := args[1].([]interface{})
  373. if len(arg1) > 0 {
  374. v1 := getFirstValidArg(arg1)
  375. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  376. if err != nil {
  377. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  378. }
  379. arg1Float64 = val
  380. }
  381. if len(arg0) > 0 {
  382. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  383. if err != nil {
  384. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  385. }
  386. deviation, err := stats.PercentileNearestRank(float64Slice, arg1Float64*100)
  387. if err != nil {
  388. if err == stats.EmptyInputErr {
  389. return nil, true
  390. }
  391. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  392. }
  393. return deviation, true
  394. }
  395. return nil, true
  396. },
  397. val: ValidateTwoNumberArg,
  398. check: returnNilIfHasAnyNil,
  399. }
  400. }
  401. func getCount(s []interface{}) int {
  402. c := 0
  403. for _, v := range s {
  404. if v != nil {
  405. c++
  406. }
  407. }
  408. return c
  409. }
  410. func getFirstValidArg(s []interface{}) interface{} {
  411. for _, v := range s {
  412. if v != nil {
  413. return v
  414. }
  415. }
  416. return nil
  417. }
  418. func sliceIntTotal(s []interface{}) (int64, error) {
  419. var total int64
  420. for _, v := range s {
  421. if v == nil {
  422. continue
  423. }
  424. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  425. if err == nil {
  426. total += vi
  427. } else if v != nil {
  428. return 0, fmt.Errorf("requires int but found %[1]T(%[1]v)", v)
  429. }
  430. }
  431. return total, nil
  432. }
  433. func sliceFloatTotal(s []interface{}) (float64, error) {
  434. var total float64
  435. for _, v := range s {
  436. if v == nil {
  437. continue
  438. }
  439. if vf, ok := v.(float64); ok {
  440. total += vf
  441. } else if v != nil {
  442. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  443. }
  444. }
  445. return total, nil
  446. }
  447. func sliceIntMax(s []interface{}, max int64) (int64, error) {
  448. for _, v := range s {
  449. if v == nil {
  450. continue
  451. }
  452. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  453. if err == nil {
  454. if vi > max {
  455. max = vi
  456. }
  457. } else if v != nil {
  458. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  459. }
  460. }
  461. return max, nil
  462. }
  463. func sliceFloatMax(s []interface{}, max float64) (float64, error) {
  464. for _, v := range s {
  465. if v == nil {
  466. continue
  467. }
  468. if vf, ok := v.(float64); ok {
  469. if max < vf {
  470. max = vf
  471. }
  472. } else if v != nil {
  473. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  474. }
  475. }
  476. return max, nil
  477. }
  478. func sliceStringMax(s []interface{}, max string) (string, error) {
  479. for _, v := range s {
  480. if v == nil {
  481. continue
  482. }
  483. if vs, ok := v.(string); ok {
  484. if max < vs {
  485. max = vs
  486. }
  487. } else if v != nil {
  488. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  489. }
  490. }
  491. return max, nil
  492. }
  493. func sliceIntMin(s []interface{}, min int64) (int64, error) {
  494. for _, v := range s {
  495. if v == nil {
  496. continue
  497. }
  498. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  499. if err == nil {
  500. if vi < min {
  501. min = vi
  502. }
  503. } else if v != nil {
  504. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  505. }
  506. }
  507. return min, nil
  508. }
  509. func sliceFloatMin(s []interface{}, min float64) (float64, error) {
  510. for _, v := range s {
  511. if v == nil {
  512. continue
  513. }
  514. if vf, ok := v.(float64); ok {
  515. if min > vf {
  516. min = vf
  517. }
  518. } else if v != nil {
  519. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  520. }
  521. }
  522. return min, nil
  523. }
  524. func sliceStringMin(s []interface{}, min string) (string, error) {
  525. for _, v := range s {
  526. if v == nil {
  527. continue
  528. }
  529. if vs, ok := v.(string); ok {
  530. if vs < min {
  531. min = vs
  532. }
  533. } else if v != nil {
  534. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  535. }
  536. }
  537. return min, nil
  538. }
  539. func dedup(r []interface{}, col []interface{}, all bool) (interface{}, error) {
  540. keyset := make(map[string]bool)
  541. result := make([]interface{}, 0)
  542. for i, m := range col {
  543. key := fmt.Sprintf("%v", m)
  544. if _, ok := keyset[key]; !ok {
  545. if all {
  546. result = append(result, r[i])
  547. } else if i == len(col)-1 {
  548. result = append(result, r[i])
  549. }
  550. keyset[key] = true
  551. }
  552. }
  553. if !all {
  554. if len(result) == 0 {
  555. return nil, nil
  556. } else {
  557. return result[0], nil
  558. }
  559. } else {
  560. return result, nil
  561. }
  562. }