funcs_agg.go 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486
  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 0, 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 fmt.Errorf("run max function error: empty data"), false
  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 fmt.Errorf("run min function error: empty data"), false
  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 0, 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. return args[0], true
  176. },
  177. val: ValidateOneArg,
  178. }
  179. builtins["deduplicate"] = builtinFunc{
  180. fType: ast.FuncTypeAgg,
  181. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  182. v1, ok1 := args[0].([]interface{})
  183. v2, ok2 := args[1].([]interface{})
  184. v3a, ok3 := args[2].([]interface{})
  185. if ok1 && ok2 && ok3 && len(v3a) > 0 {
  186. v3, ok4 := getFirstValidArg(v3a).(bool)
  187. if ok4 {
  188. if r, err := dedup(v1, v2, v3); err != nil {
  189. return err, false
  190. } else {
  191. return r, true
  192. }
  193. }
  194. }
  195. return fmt.Errorf("Invalid argument type found."), false
  196. },
  197. val: func(_ api.FunctionContext, args []ast.Expr) error {
  198. if err := ValidateLen(2, len(args)); err != nil {
  199. return err
  200. }
  201. if !ast.IsBooleanArg(args[1]) {
  202. return ProduceErrInfo(1, "bool")
  203. }
  204. return nil
  205. },
  206. }
  207. builtins["stddev"] = builtinFunc{
  208. fType: ast.FuncTypeAgg,
  209. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  210. arg0 := args[0].([]interface{})
  211. if len(arg0) > 0 {
  212. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  213. if err != nil {
  214. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  215. }
  216. deviation, err := stats.StandardDeviation(float64Slice)
  217. if err != nil {
  218. return fmt.Errorf("StandardDeviation exec with error: %v", err), false
  219. }
  220. return deviation, true
  221. }
  222. return fmt.Errorf("run stddev function error: empty data"), false
  223. },
  224. val: ValidateOneNumberArg,
  225. }
  226. builtins["stddevs"] = builtinFunc{
  227. fType: ast.FuncTypeAgg,
  228. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  229. arg0 := args[0].([]interface{})
  230. if len(arg0) > 0 {
  231. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  232. if err != nil {
  233. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  234. }
  235. deviation, err := stats.StandardDeviationSample(float64Slice)
  236. if err != nil {
  237. return fmt.Errorf("StandardDeviationSample exec with error: %v", err), false
  238. }
  239. return deviation, true
  240. }
  241. return fmt.Errorf("run stddevs function error: empty data"), false
  242. },
  243. val: ValidateOneNumberArg,
  244. }
  245. builtins["var"] = builtinFunc{
  246. fType: ast.FuncTypeAgg,
  247. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  248. arg0 := args[0].([]interface{})
  249. if len(arg0) > 0 {
  250. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  251. if err != nil {
  252. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  253. }
  254. deviation, err := stats.Variance(float64Slice)
  255. if err != nil {
  256. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  257. }
  258. return deviation, true
  259. }
  260. return fmt.Errorf("run var function error: empty data"), false
  261. },
  262. val: ValidateOneNumberArg,
  263. }
  264. builtins["vars"] = builtinFunc{
  265. fType: ast.FuncTypeAgg,
  266. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  267. arg0 := args[0].([]interface{})
  268. if len(arg0) > 0 {
  269. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  270. if err != nil {
  271. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  272. }
  273. deviation, err := stats.SampleVariance(float64Slice)
  274. if err != nil {
  275. return fmt.Errorf("SampleVariance exec with error: %v", err), false
  276. }
  277. return deviation, true
  278. }
  279. return fmt.Errorf("run vars function error: empty data"), false
  280. },
  281. val: ValidateOneNumberArg,
  282. }
  283. builtins["percentile_cont"] = builtinFunc{
  284. fType: ast.FuncTypeAgg,
  285. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  286. if err := ValidateLen(2, len(args)); err != nil {
  287. return err, false
  288. }
  289. arg0 := args[0].([]interface{})
  290. arg1 := args[1]
  291. arg1Float64, err := cast.ToFloat64(arg1, cast.CONVERT_SAMEKIND)
  292. if err != nil {
  293. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  294. }
  295. if len(arg0) > 0 {
  296. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  297. if err != nil {
  298. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  299. }
  300. deviation, err := stats.Percentile(float64Slice, arg1Float64*100)
  301. if err != nil {
  302. return fmt.Errorf("percentile exec with error: %v", err), false
  303. }
  304. return deviation, true
  305. }
  306. return fmt.Errorf("run percentile_cont function error: empty data"), false
  307. },
  308. val: ValidateTwoNumberArg,
  309. }
  310. builtins["percentile_disc"] = builtinFunc{
  311. fType: ast.FuncTypeAgg,
  312. exec: func(ctx api.FunctionContext, args []interface{}) (interface{}, bool) {
  313. if err := ValidateLen(2, len(args)); err != nil {
  314. return err, false
  315. }
  316. arg0 := args[0].([]interface{})
  317. arg1 := args[1]
  318. arg1Float64, err := cast.ToFloat64(arg1, cast.CONVERT_SAMEKIND)
  319. if err != nil {
  320. return fmt.Errorf("the second parameter requires float64 but found %[1]T(%[1]v)", arg1), false
  321. }
  322. if len(arg0) > 0 {
  323. float64Slice, err := cast.ToFloat64Slice(arg0, cast.CONVERT_SAMEKIND)
  324. if err != nil {
  325. return fmt.Errorf("requires float64 slice but found %[1]T(%[1]v)", arg0), false
  326. }
  327. deviation, err := stats.PercentileNearestRank(float64Slice, arg1Float64*100)
  328. if err != nil {
  329. return fmt.Errorf("PopulationVariance exec with error: %v", err), false
  330. }
  331. return deviation, true
  332. }
  333. return fmt.Errorf("run percentile_cont function error: empty data"), false
  334. },
  335. val: ValidateTwoNumberArg,
  336. }
  337. }
  338. func getCount(s []interface{}) int {
  339. c := 0
  340. for _, v := range s {
  341. if v != nil {
  342. c++
  343. }
  344. }
  345. return c
  346. }
  347. func getFirstValidArg(s []interface{}) interface{} {
  348. for _, v := range s {
  349. if v != nil {
  350. return v
  351. }
  352. }
  353. return nil
  354. }
  355. func sliceIntTotal(s []interface{}) (int64, error) {
  356. var total int64
  357. for _, v := range s {
  358. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  359. if err == nil {
  360. total += vi
  361. } else if v != nil {
  362. return 0, fmt.Errorf("requires int but found %[1]T(%[1]v)", v)
  363. }
  364. }
  365. return total, nil
  366. }
  367. func sliceFloatTotal(s []interface{}) (float64, error) {
  368. var total float64
  369. for _, v := range s {
  370. if vf, ok := v.(float64); ok {
  371. total += vf
  372. } else if v != nil {
  373. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  374. }
  375. }
  376. return total, nil
  377. }
  378. func sliceIntMax(s []interface{}, max int64) (int64, error) {
  379. for _, v := range s {
  380. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  381. if err == nil {
  382. if vi > max {
  383. max = vi
  384. }
  385. } else if v != nil {
  386. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  387. }
  388. }
  389. return max, nil
  390. }
  391. func sliceFloatMax(s []interface{}, max float64) (float64, error) {
  392. for _, v := range s {
  393. if vf, ok := v.(float64); ok {
  394. if max < vf {
  395. max = vf
  396. }
  397. } else if v != nil {
  398. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  399. }
  400. }
  401. return max, nil
  402. }
  403. func sliceStringMax(s []interface{}, max string) (string, error) {
  404. for _, v := range s {
  405. if vs, ok := v.(string); ok {
  406. if max < vs {
  407. max = vs
  408. }
  409. } else if v != nil {
  410. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  411. }
  412. }
  413. return max, nil
  414. }
  415. func sliceIntMin(s []interface{}, min int64) (int64, error) {
  416. for _, v := range s {
  417. vi, err := cast.ToInt64(v, cast.CONVERT_SAMEKIND)
  418. if err == nil {
  419. if vi < min {
  420. min = vi
  421. }
  422. } else if v != nil {
  423. return 0, fmt.Errorf("requires int64 but found %[1]T(%[1]v)", v)
  424. }
  425. }
  426. return min, nil
  427. }
  428. func sliceFloatMin(s []interface{}, min float64) (float64, error) {
  429. for _, v := range s {
  430. if vf, ok := v.(float64); ok {
  431. if min > vf {
  432. min = vf
  433. }
  434. } else if v != nil {
  435. return 0, fmt.Errorf("requires float64 but found %[1]T(%[1]v)", v)
  436. }
  437. }
  438. return min, nil
  439. }
  440. func sliceStringMin(s []interface{}, min string) (string, error) {
  441. for _, v := range s {
  442. if vs, ok := v.(string); ok {
  443. if vs < min {
  444. min = vs
  445. }
  446. } else if v != nil {
  447. return "", fmt.Errorf("requires string but found %[1]T(%[1]v)", v)
  448. }
  449. }
  450. return min, nil
  451. }
  452. func dedup(r []interface{}, col []interface{}, all bool) (interface{}, error) {
  453. keyset := make(map[string]bool)
  454. result := make([]interface{}, 0)
  455. for i, m := range col {
  456. key := fmt.Sprintf("%v", m)
  457. if _, ok := keyset[key]; !ok {
  458. if all {
  459. result = append(result, r[i])
  460. } else if i == len(col)-1 {
  461. result = append(result, r[i])
  462. }
  463. keyset[key] = true
  464. }
  465. }
  466. if !all {
  467. if len(result) == 0 {
  468. return nil, nil
  469. } else {
  470. return result[0], nil
  471. }
  472. } else {
  473. return result, nil
  474. }
  475. }