funcs_agg.go 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560
  1. // Copyright 2022 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["deduplicate"] = builtinFunc{
  188. fType: ast.FuncTypeAgg,
  189. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  190. v1, ok1 := args[0].([]interface{})
  191. v2, ok2 := args[1].([]interface{})
  192. v3a, ok3 := args[2].([]interface{})
  193. if ok1 && ok2 && ok3 && len(v3a) > 0 {
  194. v3, ok4 := getFirstValidArg(v3a).(bool)
  195. if ok4 {
  196. if r, err := dedup(v1, v2, v3); err != nil {
  197. return err, false
  198. } else {
  199. return r, true
  200. }
  201. }
  202. }
  203. return fmt.Errorf("Invalid argument type found."), false
  204. },
  205. val: func(_ api.FunctionContext, args []ast.Expr) error {
  206. if err := ValidateLen(2, len(args)); err != nil {
  207. return err
  208. }
  209. if !ast.IsBooleanArg(args[1]) {
  210. return ProduceErrInfo(1, "bool")
  211. }
  212. return nil
  213. },
  214. check: returnNilIfHasAnyNil,
  215. }
  216. builtins["stddev"] = builtinFunc{
  217. fType: ast.FuncTypeAgg,
  218. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  219. arg0 := args[0].([]interface{})
  220. if len(arg0) > 0 {
  221. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  222. if err != nil {
  223. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  224. }
  225. deviation, err := stats.StandardDeviation(float64Slice)
  226. if err != nil {
  227. if err == stats.EmptyInputErr {
  228. return nil, true
  229. }
  230. return fmt.Errorf("StandardDeviation exec with error: %v", err), false
  231. }
  232. return deviation, true
  233. }
  234. return nil, true
  235. },
  236. val: ValidateOneNumberArg,
  237. check: returnNilIfHasAnyNil,
  238. }
  239. builtins["stddevs"] = builtinFunc{
  240. fType: ast.FuncTypeAgg,
  241. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  242. arg0 := args[0].([]interface{})
  243. if len(arg0) > 0 {
  244. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  245. if err != nil {
  246. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  247. }
  248. deviation, err := stats.StandardDeviationSample(float64Slice)
  249. if err != nil {
  250. if err == stats.EmptyInputErr {
  251. return nil, true
  252. }
  253. return fmt.Errorf("StandardDeviationSample exec with error: %v", err), false
  254. }
  255. return deviation, true
  256. }
  257. return nil, true
  258. },
  259. val: ValidateOneNumberArg,
  260. check: returnNilIfHasAnyNil,
  261. }
  262. builtins["var"] = builtinFunc{
  263. fType: ast.FuncTypeAgg,
  264. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  265. arg0 := args[0].([]interface{})
  266. if len(arg0) > 0 {
  267. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  268. if err != nil {
  269. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  270. }
  271. deviation, err := stats.Variance(float64Slice)
  272. if err != nil {
  273. if err == stats.EmptyInputErr {
  274. return nil, true
  275. }
  276. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  277. }
  278. return deviation, true
  279. }
  280. return nil, true
  281. },
  282. val: ValidateOneNumberArg,
  283. check: returnNilIfHasAnyNil,
  284. }
  285. builtins["vars"] = builtinFunc{
  286. fType: ast.FuncTypeAgg,
  287. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  288. arg0 := args[0].([]interface{})
  289. if len(arg0) > 0 {
  290. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  291. if err != nil {
  292. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  293. }
  294. deviation, err := stats.SampleVariance(float64Slice)
  295. if err != nil {
  296. if err == stats.EmptyInputErr {
  297. return nil, true
  298. }
  299. return fmt.Errorf("SampleVariance exec with error: %v", err), false
  300. }
  301. return deviation, true
  302. }
  303. return nil, true
  304. },
  305. val: ValidateOneNumberArg,
  306. check: returnNilIfHasAnyNil,
  307. }
  308. builtins["percentile_cont"] = builtinFunc{
  309. fType: ast.FuncTypeAgg,
  310. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  311. if err := ValidateLen(2, len(args)); err != nil {
  312. return err, false
  313. }
  314. var arg1Float64 float64 = 1
  315. arg0 := args[0].([]interface{})
  316. arg1 := args[1].([]interface{})
  317. if len(arg1) > 0 {
  318. v1 := getFirstValidArg(arg1)
  319. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  320. if err != nil {
  321. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  322. }
  323. arg1Float64 = val
  324. }
  325. if len(arg0) > 0 {
  326. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  327. if err != nil {
  328. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  329. }
  330. deviation, err := stats.Percentile(float64Slice, arg1Float64*100)
  331. if err != nil {
  332. if err == stats.EmptyInputErr {
  333. return nil, true
  334. }
  335. return fmt.Errorf("percentile exec with error: %v", err), false
  336. }
  337. return deviation, true
  338. }
  339. return nil, true
  340. },
  341. val: ValidateTwoNumberArg,
  342. check: returnNilIfHasAnyNil,
  343. }
  344. builtins["percentile_disc"] = builtinFunc{
  345. fType: ast.FuncTypeAgg,
  346. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  347. if err := ValidateLen(2, len(args)); err != nil {
  348. return err, false
  349. }
  350. var arg1Float64 float64 = 1
  351. arg0 := args[0].([]interface{})
  352. arg1 := args[1].([]interface{})
  353. if len(arg1) > 0 {
  354. v1 := getFirstValidArg(arg1)
  355. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  356. if err != nil {
  357. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  358. }
  359. arg1Float64 = val
  360. }
  361. if len(arg0) > 0 {
  362. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  363. if err != nil {
  364. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  365. }
  366. deviation, err := stats.PercentileNearestRank(float64Slice, arg1Float64*100)
  367. if err != nil {
  368. if err == stats.EmptyInputErr {
  369. return nil, true
  370. }
  371. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  372. }
  373. return deviation, true
  374. }
  375. return nil, true
  376. },
  377. val: ValidateTwoNumberArg,
  378. check: returnNilIfHasAnyNil,
  379. }
  380. }
  381. func getCount(s []interface{}) int {
  382. c := 0
  383. for _, v := range s {
  384. if v != nil {
  385. c++
  386. }
  387. }
  388. return c
  389. }
  390. func getFirstValidArg(s []interface{}) interface{} {
  391. for _, v := range s {
  392. if v != nil {
  393. return v
  394. }
  395. }
  396. return nil
  397. }
  398. func sliceIntTotal(s []interface{}) (int64, error) {
  399. var total int64
  400. for _, v := range s {
  401. if v == nil {
  402. continue
  403. }
  404. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  405. if err == nil {
  406. total += vi
  407. } else if v != nil {
  408. return 0, fmt.Errorf("requires int but found %[1]T(%[1]v)", v)
  409. }
  410. }
  411. return total, nil
  412. }
  413. func sliceFloatTotal(s []interface{}) (float64, error) {
  414. var total float64
  415. for _, v := range s {
  416. if v == nil {
  417. continue
  418. }
  419. if vf, ok := v.(float64); ok {
  420. total += vf
  421. } else if v != nil {
  422. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  423. }
  424. }
  425. return total, nil
  426. }
  427. func sliceIntMax(s []interface{}, max int64) (int64, error) {
  428. for _, v := range s {
  429. if v == nil {
  430. continue
  431. }
  432. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  433. if err == nil {
  434. if vi > max {
  435. max = vi
  436. }
  437. } else if v != nil {
  438. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  439. }
  440. }
  441. return max, nil
  442. }
  443. func sliceFloatMax(s []interface{}, max float64) (float64, error) {
  444. for _, v := range s {
  445. if v == nil {
  446. continue
  447. }
  448. if vf, ok := v.(float64); ok {
  449. if max < vf {
  450. max = vf
  451. }
  452. } else if v != nil {
  453. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  454. }
  455. }
  456. return max, nil
  457. }
  458. func sliceStringMax(s []interface{}, max string) (string, error) {
  459. for _, v := range s {
  460. if v == nil {
  461. continue
  462. }
  463. if vs, ok := v.(string); ok {
  464. if max < vs {
  465. max = vs
  466. }
  467. } else if v != nil {
  468. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  469. }
  470. }
  471. return max, nil
  472. }
  473. func sliceIntMin(s []interface{}, min int64) (int64, error) {
  474. for _, v := range s {
  475. if v == nil {
  476. continue
  477. }
  478. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  479. if err == nil {
  480. if vi < min {
  481. min = vi
  482. }
  483. } else if v != nil {
  484. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  485. }
  486. }
  487. return min, nil
  488. }
  489. func sliceFloatMin(s []interface{}, min float64) (float64, error) {
  490. for _, v := range s {
  491. if v == nil {
  492. continue
  493. }
  494. if vf, ok := v.(float64); ok {
  495. if min > vf {
  496. min = vf
  497. }
  498. } else if v != nil {
  499. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  500. }
  501. }
  502. return min, nil
  503. }
  504. func sliceStringMin(s []interface{}, min string) (string, error) {
  505. for _, v := range s {
  506. if v == nil {
  507. continue
  508. }
  509. if vs, ok := v.(string); ok {
  510. if vs < min {
  511. min = vs
  512. }
  513. } else if v != nil {
  514. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  515. }
  516. }
  517. return min, nil
  518. }
  519. func dedup(r []interface{}, col []interface{}, all bool) (interface{}, error) {
  520. keyset := make(map[string]bool)
  521. result := make([]interface{}, 0)
  522. for i, m := range col {
  523. key := fmt.Sprintf("%v", m)
  524. if _, ok := keyset[key]; !ok {
  525. if all {
  526. result = append(result, r[i])
  527. } else if i == len(col)-1 {
  528. result = append(result, r[i])
  529. }
  530. keyset[key] = true
  531. }
  532. }
  533. if !all {
  534. if len(result) == 0 {
  535. return nil, nil
  536. } else {
  537. return result[0], nil
  538. }
  539. } else {
  540. return result, nil
  541. }
  542. }