funcs_agg.go 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519
  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/lf-edge/ekuiper/pkg/api"
  18. "github.com/lf-edge/ekuiper/pkg/ast"
  19. "github.com/lf-edge/ekuiper/pkg/cast"
  20. "github.com/montanaflynn/stats"
  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. }
  53. builtins["count"] = builtinFunc{
  54. fType: ast.FuncTypeAgg,
  55. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  56. arg0 := args[0].([]interface{})
  57. return getCount(arg0), true
  58. },
  59. val: ValidateOneArg,
  60. }
  61. builtins["max"] = builtinFunc{
  62. fType: ast.FuncTypeAgg,
  63. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  64. arg0 := args[0].([]interface{})
  65. if len(arg0) > 0 {
  66. v := getFirstValidArg(arg0)
  67. switch t := v.(type) {
  68. case int:
  69. if r, err := sliceIntMax(arg0, int64(t)); err != nil {
  70. return err, false
  71. } else {
  72. return r, true
  73. }
  74. case int64:
  75. if r, err := sliceIntMax(arg0, t); err != nil {
  76. return err, false
  77. } else {
  78. return r, true
  79. }
  80. case float64:
  81. if r, err := sliceFloatMax(arg0, t); err != nil {
  82. return err, false
  83. } else {
  84. return r, true
  85. }
  86. case string:
  87. if r, err := sliceStringMax(arg0, t); err != nil {
  88. return err, false
  89. } else {
  90. return r, true
  91. }
  92. case nil:
  93. return nil, true
  94. default:
  95. return fmt.Errorf("run max function error: found invalid arg %[1]T(%[1]v)", v), false
  96. }
  97. }
  98. return nil, true
  99. },
  100. val: ValidateOneNumberArg,
  101. }
  102. builtins["min"] = builtinFunc{
  103. fType: ast.FuncTypeAgg,
  104. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  105. arg0 := args[0].([]interface{})
  106. if len(arg0) > 0 {
  107. v := getFirstValidArg(arg0)
  108. switch t := v.(type) {
  109. case int:
  110. if r, err := sliceIntMin(arg0, int64(t)); err != nil {
  111. return err, false
  112. } else {
  113. return r, true
  114. }
  115. case int64:
  116. if r, err := sliceIntMin(arg0, t); err != nil {
  117. return err, false
  118. } else {
  119. return r, true
  120. }
  121. case float64:
  122. if r, err := sliceFloatMin(arg0, t); err != nil {
  123. return err, false
  124. } else {
  125. return r, true
  126. }
  127. case string:
  128. if r, err := sliceStringMin(arg0, t); err != nil {
  129. return err, false
  130. } else {
  131. return r, true
  132. }
  133. case nil:
  134. return nil, true
  135. default:
  136. return fmt.Errorf("run min function error: found invalid arg %[1]T(%[1]v)", v), false
  137. }
  138. }
  139. return nil, true
  140. },
  141. val: ValidateOneNumberArg,
  142. }
  143. builtins["sum"] = builtinFunc{
  144. fType: ast.FuncTypeAgg,
  145. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  146. arg0 := args[0].([]interface{})
  147. if len(arg0) > 0 {
  148. v := getFirstValidArg(arg0)
  149. switch v.(type) {
  150. case int, int64:
  151. if r, err := sliceIntTotal(arg0); err != nil {
  152. return err, false
  153. } else {
  154. return r, true
  155. }
  156. case float64:
  157. if r, err := sliceFloatTotal(arg0); err != nil {
  158. return err, false
  159. } else {
  160. return r, true
  161. }
  162. case nil:
  163. return nil, true
  164. default:
  165. return fmt.Errorf("run sum function error: found invalid arg %[1]T(%[1]v)", v), false
  166. }
  167. }
  168. return nil, true
  169. },
  170. val: ValidateOneNumberArg,
  171. }
  172. builtins["collect"] = builtinFunc{
  173. fType: ast.FuncTypeAgg,
  174. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  175. if len(args) > 0 {
  176. return args[0], true
  177. }
  178. return make([]interface{}, 0), true
  179. },
  180. val: ValidateOneArg,
  181. }
  182. builtins["deduplicate"] = builtinFunc{
  183. fType: ast.FuncTypeAgg,
  184. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  185. v1, ok1 := args[0].([]interface{})
  186. v2, ok2 := args[1].([]interface{})
  187. v3a, ok3 := args[2].([]interface{})
  188. if ok1 && ok2 && ok3 && len(v3a) > 0 {
  189. v3, ok4 := getFirstValidArg(v3a).(bool)
  190. if ok4 {
  191. if r, err := dedup(v1, v2, v3); err != nil {
  192. return err, false
  193. } else {
  194. return r, true
  195. }
  196. }
  197. }
  198. return fmt.Errorf("Invalid argument type found."), false
  199. },
  200. val: func(_ api.FunctionContext, args []ast.Expr) error {
  201. if err := ValidateLen(2, len(args)); err != nil {
  202. return err
  203. }
  204. if !ast.IsBooleanArg(args[1]) {
  205. return ProduceErrInfo(1, "bool")
  206. }
  207. return nil
  208. },
  209. }
  210. builtins["stddev"] = builtinFunc{
  211. fType: ast.FuncTypeAgg,
  212. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  213. arg0 := args[0].([]interface{})
  214. if len(arg0) > 0 {
  215. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  216. if err != nil {
  217. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  218. }
  219. deviation, err := stats.StandardDeviation(float64Slice)
  220. if err != nil {
  221. if err == stats.EmptyInputErr {
  222. return nil, true
  223. }
  224. return fmt.Errorf("StandardDeviation exec with error: %v", err), false
  225. }
  226. return deviation, true
  227. }
  228. return nil, true
  229. },
  230. val: ValidateOneNumberArg,
  231. }
  232. builtins["stddevs"] = builtinFunc{
  233. fType: ast.FuncTypeAgg,
  234. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  235. arg0 := args[0].([]interface{})
  236. if len(arg0) > 0 {
  237. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  238. if err != nil {
  239. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  240. }
  241. deviation, err := stats.StandardDeviationSample(float64Slice)
  242. if err != nil {
  243. if err == stats.EmptyInputErr {
  244. return nil, true
  245. }
  246. return fmt.Errorf("StandardDeviationSample exec with error: %v", err), false
  247. }
  248. return deviation, true
  249. }
  250. return nil, true
  251. },
  252. val: ValidateOneNumberArg,
  253. }
  254. builtins["var"] = builtinFunc{
  255. fType: ast.FuncTypeAgg,
  256. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  257. arg0 := args[0].([]interface{})
  258. if len(arg0) > 0 {
  259. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  260. if err != nil {
  261. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  262. }
  263. deviation, err := stats.Variance(float64Slice)
  264. if err != nil {
  265. if err == stats.EmptyInputErr {
  266. return nil, true
  267. }
  268. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  269. }
  270. return deviation, true
  271. }
  272. return nil, true
  273. },
  274. val: ValidateOneNumberArg,
  275. }
  276. builtins["vars"] = builtinFunc{
  277. fType: ast.FuncTypeAgg,
  278. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  279. arg0 := args[0].([]interface{})
  280. if len(arg0) > 0 {
  281. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  282. if err != nil {
  283. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  284. }
  285. deviation, err := stats.SampleVariance(float64Slice)
  286. if err != nil {
  287. if err == stats.EmptyInputErr {
  288. return nil, true
  289. }
  290. return fmt.Errorf("SampleVariance exec with error: %v", err), false
  291. }
  292. return deviation, true
  293. }
  294. return nil, true
  295. },
  296. val: ValidateOneNumberArg,
  297. }
  298. builtins["percentile_cont"] = builtinFunc{
  299. fType: ast.FuncTypeAgg,
  300. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  301. if err := ValidateLen(2, len(args)); err != nil {
  302. return err, false
  303. }
  304. var arg1Float64 float64 = 1
  305. arg0 := args[0].([]interface{})
  306. arg1 := args[1].([]interface{})
  307. if len(arg1) > 0 {
  308. v1 := getFirstValidArg(arg1)
  309. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  310. if err != nil {
  311. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  312. }
  313. arg1Float64 = val
  314. }
  315. if len(arg0) > 0 {
  316. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  317. if err != nil {
  318. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  319. }
  320. deviation, err := stats.Percentile(float64Slice, arg1Float64*100)
  321. if err != nil {
  322. if err == stats.EmptyInputErr {
  323. return nil, true
  324. }
  325. return fmt.Errorf("percentile exec with error: %v", err), false
  326. }
  327. return deviation, true
  328. }
  329. return nil, true
  330. },
  331. val: ValidateTwoNumberArg,
  332. }
  333. builtins["percentile_disc"] = builtinFunc{
  334. fType: ast.FuncTypeAgg,
  335. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  336. if err := ValidateLen(2, len(args)); err != nil {
  337. return err, false
  338. }
  339. var arg1Float64 float64 = 1
  340. arg0 := args[0].([]interface{})
  341. arg1 := args[1].([]interface{})
  342. if len(arg1) > 0 {
  343. v1 := getFirstValidArg(arg1)
  344. val, err := cast.ToFloat64(v1, cast.CONVERT_SAMEKIND)
  345. if err != nil {
  346. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  347. }
  348. arg1Float64 = val
  349. }
  350. if len(arg0) > 0 {
  351. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  352. if err != nil {
  353. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  354. }
  355. deviation, err := stats.PercentileNearestRank(float64Slice, arg1Float64*100)
  356. if err != nil {
  357. if err == stats.EmptyInputErr {
  358. return nil, true
  359. }
  360. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  361. }
  362. return deviation, true
  363. }
  364. return nil, true
  365. },
  366. val: ValidateTwoNumberArg,
  367. }
  368. }
  369. func getCount(s []interface{}) int {
  370. c := 0
  371. for _, v := range s {
  372. if v != nil {
  373. c++
  374. }
  375. }
  376. return c
  377. }
  378. func getFirstValidArg(s []interface{}) interface{} {
  379. for _, v := range s {
  380. if v != nil {
  381. return v
  382. }
  383. }
  384. return nil
  385. }
  386. func sliceIntTotal(s []interface{}) (int64, error) {
  387. var total int64
  388. for _, v := range s {
  389. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  390. if err == nil {
  391. total += vi
  392. } else if v != nil {
  393. return 0, fmt.Errorf("requires int but found %[1]T(%[1]v)", v)
  394. }
  395. }
  396. return total, nil
  397. }
  398. func sliceFloatTotal(s []interface{}) (float64, error) {
  399. var total float64
  400. for _, v := range s {
  401. if vf, ok := v.(float64); ok {
  402. total += vf
  403. } else if v != nil {
  404. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  405. }
  406. }
  407. return total, nil
  408. }
  409. func sliceIntMax(s []interface{}, max int64) (int64, error) {
  410. for _, v := range s {
  411. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  412. if err == nil {
  413. if vi > max {
  414. max = vi
  415. }
  416. } else if v != nil {
  417. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  418. }
  419. }
  420. return max, nil
  421. }
  422. func sliceFloatMax(s []interface{}, max float64) (float64, error) {
  423. for _, v := range s {
  424. if vf, ok := v.(float64); ok {
  425. if max < vf {
  426. max = vf
  427. }
  428. } else if v != nil {
  429. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  430. }
  431. }
  432. return max, nil
  433. }
  434. func sliceStringMax(s []interface{}, max string) (string, error) {
  435. for _, v := range s {
  436. if vs, ok := v.(string); ok {
  437. if max < vs {
  438. max = vs
  439. }
  440. } else if v != nil {
  441. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  442. }
  443. }
  444. return max, nil
  445. }
  446. func sliceIntMin(s []interface{}, min int64) (int64, error) {
  447. for _, v := range s {
  448. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  449. if err == nil {
  450. if vi < min {
  451. min = vi
  452. }
  453. } else if v != nil {
  454. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  455. }
  456. }
  457. return min, nil
  458. }
  459. func sliceFloatMin(s []interface{}, min float64) (float64, error) {
  460. for _, v := range s {
  461. if vf, ok := v.(float64); ok {
  462. if min > vf {
  463. min = vf
  464. }
  465. } else if v != nil {
  466. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  467. }
  468. }
  469. return min, nil
  470. }
  471. func sliceStringMin(s []interface{}, min string) (string, error) {
  472. for _, v := range s {
  473. if vs, ok := v.(string); ok {
  474. if vs < min {
  475. min = vs
  476. }
  477. } else if v != nil {
  478. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  479. }
  480. }
  481. return min, nil
  482. }
  483. func dedup(r []interface{}, col []interface{}, all bool) (interface{}, error) {
  484. keyset := make(map[string]bool)
  485. result := make([]interface{}, 0)
  486. for i, m := range col {
  487. key := fmt.Sprintf("%v", m)
  488. if _, ok := keyset[key]; !ok {
  489. if all {
  490. result = append(result, r[i])
  491. } else if i == len(col)-1 {
  492. result = append(result, r[i])
  493. }
  494. keyset[key] = true
  495. }
  496. }
  497. if !all {
  498. if len(result) == 0 {
  499. return nil, nil
  500. } else {
  501. return result[0], nil
  502. }
  503. } else {
  504. return result, nil
  505. }
  506. }