From dd875ac76df9d0fdde19b632ebdce26be41066f9 Mon Sep 17 00:00:00 2001 From: venom1204 Date: Sat, 18 Jul 2026 10:10:27 +0000 Subject: [PATCH] updated changes --- R/shift.R | 12 ++++-- inst/tests/tests.Rraw | 6 +++ man/nafill.Rd | 11 +++++- src/data.table.h | 8 ++-- src/nafill.c | 88 +++++++++++++++++++++++++++++++++---------- 5 files changed, 97 insertions(+), 28 deletions(-) diff --git a/R/shift.R b/R/shift.R index 1c68d13c41..c586bed0b6 100644 --- a/R/shift.R +++ b/R/shift.R @@ -26,12 +26,16 @@ shift = function(x, n=1L, fill, type=c("lag", "lead", "shift", "cyclic"), give.n ans } -nafill = function(x, type=c("const","locf","nocb"), fill=NA, nan=NA) { +nafill = function(x, type=c("const","locf","nocb"), fill=NA, nan=NA, limit=Inf) { type = match.arg(type) - .Call(CnafillR, x, type, fill, nan_is_na(nan), FALSE, NULL) + if (!is.numeric(limit) || length(limit) != 1L || limit < 0) + stopf("limit must be a non-negative scalar numeric or Inf") + .Call(CnafillR, x, type, fill, nan_is_na(nan), FALSE, NULL, as.double(limit)) } -setnafill = function(x, type=c("const","locf","nocb"), fill=NA, nan=NA, cols=seq_along(x)) { +setnafill = function(x, type=c("const","locf","nocb"), fill=NA, nan=NA, cols=seq_along(x), limit=Inf) { type = match.arg(type) - invisible(.Call(CnafillR, x, type, fill, nan_is_na(nan), TRUE, cols)) + if (!is.numeric(limit) || length(limit) != 1L || limit < 0) + stopf("limit must be a non-negative scalar numeric or Inf") + invisible(.Call(CnafillR, x, type, fill, nan_is_na(nan), TRUE, cols, as.double(limit))) } diff --git a/inst/tests/tests.Rraw b/inst/tests/tests.Rraw index 00973ae3a5..5c72eca1de 100644 --- a/inst/tests/tests.Rraw +++ b/inst/tests/tests.Rraw @@ -21831,3 +21831,9 @@ test(2378.92, .Call(CresizeVector, x, 2:3), error = "must be length 1 non-NA non test(2378.93, .Call(CresizeVector, x, -2L), error = "must be length 1 non-NA non-negative integer value") test(2378.94, .Call(CresizeVector, x, NA_integer_), error = "must be length 1 non-NA non-negative integer value") test(2378.95, .Call(CresizeVector, NULL, 2L), error = "must be a vector") + +# #7677: add limit argument to nafill() +test(2379.01, nafill(c(1, NA, NA, NA, 5), type="locf", limit=1), c(1, 1, NA, NA, 5)) +test(2379.02, nafill(c(1, NA, NA, 5, NA, NA, NA, 9), type="locf", limit=2), c(1, 1, 1, 5, 5, 5, NA, 9)) +test(2379.03, nafill(c(1, NA, NA, 4, NA), type="locf", limit=Inf), c(1, 1, 1, 4, 4)) +test(2379.04, nafill(c(NA, NA, 3, NA, NA), type="locf", limit=1), c(NA, NA, 3, 3, NA)) diff --git a/man/nafill.Rd b/man/nafill.Rd index 90c4b1c5c6..57c2742b8e 100644 --- a/man/nafill.Rd +++ b/man/nafill.Rd @@ -10,8 +10,8 @@ Fast fill missing values using constant value, \emph{last observation carried forward} or \emph{next observation carried backward}. } \usage{ -nafill(x, type=c("const", "locf", "nocb"), fill=NA, nan=NA) -setnafill(x, type=c("const", "locf", "nocb"), fill=NA, nan=NA, cols=seq_along(x)) +nafill(x, type=c("const", "locf", "nocb"), fill=NA, nan=NA, limit=Inf) +setnafill(x, type=c("const", "locf", "nocb"), fill=NA, nan=NA, cols=seq_along(x), limit=Inf) } \arguments{ \item{x}{ Vector, list, data.frame or data.table of logical, numeric or character columns. } @@ -19,6 +19,7 @@ setnafill(x, type=c("const", "locf", "nocb"), fill=NA, nan=NA, cols=seq_along(x) \item{fill}{ Value to be used to replace missing observations. See examples. } \item{nan}{ Either \code{NaN} or \code{NA}; if the former, \code{NaN} is treated as distinct from \code{NA}, otherwise, they are treated the same during replacement. See Examples. } \item{cols}{ Numeric or character vector specifying columns to be updated. } + \item{limit}{ The maximum number of consecutive \code{NA} values to fill. Must be a non-negative scalar numeric. Default is \code{Inf}. } } \details{ Supported types are \emph{logical}, \emph{integer}, \emph{double}, \emph{character}, and \emph{factor}, as well as classes built on top of these such as \code{Date}, \code{IDate}, and \code{POSIXct}. @@ -51,6 +52,12 @@ nafill(dt, "nocb") setnafill(dt, "locf", cols=c("v2","v3")) dt + +# limit= restricts the number of consecutive fills +y = c(1, NA, NA, NA, 5) +nafill(y, "locf", limit=1) # Only fills the first NA +nafill(y, "locf", limit=2) # Fills the first two NAs + } \seealso{ \code{\link{shift}}, \code{\link{data.table}}, \code{\link{fcoalesce}} diff --git a/src/data.table.h b/src/data.table.h index dbfe495edb..9427ff7745 100644 --- a/src/data.table.h +++ b/src/data.table.h @@ -311,9 +311,11 @@ SEXP copyAsGrowable(SEXP x); SEXP resizeVector(SEXP x, SEXP size); // nafill.c -void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, bool nan_is_na, ans_t *ans, bool verbose); -void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill, ans_t *ans, bool verbose); -SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, SEXP cols); +void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, bool nan_is_na, ans_t *ans, bool verbose, double limit); +void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill, ans_t *ans, bool verbose, double limit); +void nafillInteger64(int64_t *x, uint_fast64_t nx, unsigned int type, int64_t fill, ans_t *ans, bool verbose, double limit); +void nafillString(const SEXP *x, uint_fast64_t nx, unsigned int type, SEXP fill, ans_t *ans, bool verbose, double limit); +SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, SEXP cols, SEXP limit); // between.c SEXP between(SEXP x, SEXP lower, SEXP upper, SEXP incbounds, SEXP NAbounds, SEXP check); diff --git a/src/nafill.c b/src/nafill.c index 8a8e8eab56..197fe954cb 100644 --- a/src/nafill.c +++ b/src/nafill.c @@ -1,6 +1,6 @@ #include "data.table.h" -void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, bool nan_is_na, ans_t *ans, bool verbose) { +void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, bool nan_is_na, ans_t *ans, bool verbose, double limit) { double tic=0.0; if (verbose) tic = omp_get_wtime(); @@ -15,34 +15,52 @@ void nafillDouble(double *x, uint_fast64_t nx, unsigned int type, double fill, b } } } else if (type==1) { // locf + uint_fast64_t fills = 0; if (nan_is_na) { ans->dbl_v[0] = ISNAN(x[0]) ? fill : x[0]; + if (ISNAN(x[0])) fills = 1; for (uint_fast64_t i=1; idbl_v[i] = ISNAN(x[i]) ? ans->dbl_v[i-1] : x[i]; + if (ISNAN(x[i])) { + if (fills < limit) { ans->dbl_v[i] = ans->dbl_v[i-1]; fills++; } + else ans->dbl_v[i] = x[i]; + } else { ans->dbl_v[i] = x[i]; fills = 0; } } } else { ans->dbl_v[0] = ISNA(x[0]) ? fill : x[0]; + if (ISNA(x[0])) fills = 1; for (uint_fast64_t i=1; idbl_v[i] = ISNA(x[i]) ? ans->dbl_v[i-1] : x[i]; + if (ISNA(x[i])) { + if (fills < limit) { ans->dbl_v[i] = ans->dbl_v[i-1]; fills++; } + else ans->dbl_v[i] = x[i]; + } else { ans->dbl_v[i] = x[i]; fills = 0; } } } } else if (type==2) { // nocb + uint_fast64_t fills = 0; if (nan_is_na) { ans->dbl_v[nx-1] = ISNAN(x[nx-1]) ? fill : x[nx-1]; + if (ISNAN(x[nx-1])) fills = 1; for (int_fast64_t i=nx-2; i>=0; i--) { - ans->dbl_v[i] = ISNAN(x[i]) ? ans->dbl_v[i+1] : x[i]; + if (ISNAN(x[i])) { + if (fills < limit) { ans->dbl_v[i] = ans->dbl_v[i+1]; fills++; } + else ans->dbl_v[i] = x[i]; + } else { ans->dbl_v[i] = x[i]; fills = 0; } } } else { ans->dbl_v[nx-1] = ISNA(x[nx-1]) ? fill : x[nx-1]; + if (ISNA(x[nx-1])) fills = 1; for (int_fast64_t i=nx-2; i>=0; i--) { - ans->dbl_v[i] = ISNA(x[i]) ? ans->dbl_v[i+1] : x[i]; + if (ISNA(x[i])) { + if (fills < limit) { ans->dbl_v[i] = ans->dbl_v[i+1]; fills++; } + else ans->dbl_v[i] = x[i]; + } else { ans->dbl_v[i] = x[i]; fills = 0; } } } } if (verbose) snprintf(ans->message[0], 500, _("%s: took %.3fs\n"), __func__, omp_get_wtime()-tic); } -void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill, ans_t *ans, bool verbose) { +void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill, ans_t *ans, bool verbose, double limit) { double tic=0.0; if (verbose) tic = omp_get_wtime(); @@ -51,20 +69,30 @@ void nafillInteger(int32_t *x, uint_fast64_t nx, unsigned int type, int32_t fill ans->int_v[i] = x[i]==NA_INTEGER ? fill : x[i]; } } else if (type==1) { // locf + uint_fast64_t fills = 0; ans->int_v[0] = x[0]==NA_INTEGER ? fill : x[0]; + if (x[0]==NA_INTEGER) fills = 1; for (uint_fast64_t i=1; iint_v[i] = x[i]==NA_INTEGER ? ans->int_v[i-1] : x[i]; + if (x[i]==NA_INTEGER) { + if (fills < limit) { ans->int_v[i] = ans->int_v[i-1]; fills++; } + else ans->int_v[i] = x[i]; + } else { ans->int_v[i] = x[i]; fills = 0; } } } else if (type==2) { // nocb + uint_fast64_t fills = 0; ans->int_v[nx-1] = x[nx-1]==NA_INTEGER ? fill : x[nx-1]; + if (x[nx-1]==NA_INTEGER) fills = 1; for (int_fast64_t i=nx-2; i>=0; i--) { - ans->int_v[i] = x[i]==NA_INTEGER ? ans->int_v[i+1] : x[i]; + if (x[i]==NA_INTEGER) { + if (fills < limit) { ans->int_v[i] = ans->int_v[i+1]; fills++; } + else ans->int_v[i] = x[i]; + } else { ans->int_v[i] = x[i]; fills = 0; } } } if (verbose) snprintf(ans->message[0], 500, _("%s: took %.3fs\n"), __func__, omp_get_wtime()-tic); } -void nafillInteger64(int64_t *x, uint_fast64_t nx, unsigned int type, int64_t fill, ans_t *ans, bool verbose) { +void nafillInteger64(int64_t *x, uint_fast64_t nx, unsigned int type, int64_t fill, ans_t *ans, bool verbose, double limit) { double tic=0.0; if (verbose) tic = omp_get_wtime(); @@ -73,21 +101,31 @@ void nafillInteger64(int64_t *x, uint_fast64_t nx, unsigned int type, int64_t fi ans->int64_v[i] = x[i]==NA_INTEGER64 ? fill : x[i]; } } else if (type==1) { // locf + uint_fast64_t fills = 0; ans->int64_v[0] = x[0]==NA_INTEGER64 ? fill : x[0]; + if (x[0]==NA_INTEGER64) fills = 1; for (uint_fast64_t i=1; iint64_v[i] = x[i]==NA_INTEGER64 ? ans->int64_v[i-1] : x[i]; + if (x[i]==NA_INTEGER64) { + if (fills < limit) { ans->int64_v[i] = ans->int64_v[i-1]; fills++; } + else ans->int64_v[i] = x[i]; + } else { ans->int64_v[i] = x[i]; fills = 0; } } } else if (type==2) { // nocb + uint_fast64_t fills = 0; ans->int64_v[nx-1] = x[nx-1]==NA_INTEGER64 ? fill : x[nx-1]; + if (x[nx-1]==NA_INTEGER64) fills = 1; for (int_fast64_t i=nx-2; i>=0; i--) { - ans->int64_v[i] = x[i]==NA_INTEGER64 ? ans->int64_v[i+1] : x[i]; + if (x[i]==NA_INTEGER64) { + if (fills < limit) { ans->int64_v[i] = ans->int64_v[i+1]; fills++; } + else ans->int64_v[i] = x[i]; + } else { ans->int64_v[i] = x[i]; fills = 0; } } } if (verbose) snprintf(ans->message[0], 500, _("%s: took %.3fs\n"), __func__, omp_get_wtime()-tic); } -void nafillString(const SEXP *x, uint_fast64_t nx, unsigned int type, SEXP fill, ans_t *ans, bool verbose) { +void nafillString(const SEXP *x, uint_fast64_t nx, unsigned int type, SEXP fill, ans_t *ans, bool verbose, double limit) { double tic=0.0; if (verbose) tic = omp_get_wtime(); @@ -96,16 +134,26 @@ void nafillString(const SEXP *x, uint_fast64_t nx, unsigned int type, SEXP fill, SET_STRING_ELT(ans->char_v, i, x[i]==NA_STRING ? fill : x[i]); } } else if (type==1) { // locf + uint_fast64_t fills = 0; SET_STRING_ELT(ans->char_v, 0, x[0]==NA_STRING ? fill : x[0]); + if (x[0]==NA_STRING) fills = 1; const SEXP* thisans = SEXPPTR_RO(ans->char_v); // takes out STRING_ELT from loop for (uint_fast64_t i=1; ichar_v, i, x[i]==NA_STRING ? thisans[i-1] : x[i]); + if (x[i]==NA_STRING) { + if (fills < limit) { SET_STRING_ELT(ans->char_v, i, thisans[i-1]); fills++; } + else SET_STRING_ELT(ans->char_v, i, x[i]); + } else { SET_STRING_ELT(ans->char_v, i, x[i]); fills = 0; } } } else if (type==2) { // nocb + uint_fast64_t fills = 0; SET_STRING_ELT(ans->char_v, nx-1, x[nx-1]==NA_STRING ? fill : x[nx-1]); + if (x[nx-1]==NA_STRING) fills = 1; const SEXP* thisans = SEXPPTR_RO(ans->char_v); // takes out STRING_ELT from loop for (int_fast64_t i=nx-2; i>=0; i--) { - SET_STRING_ELT(ans->char_v, i, x[i]==NA_STRING ? thisans[i+1] : x[i]); + if (x[i]==NA_STRING) { + if (fills < limit) { SET_STRING_ELT(ans->char_v, i, thisans[i+1]); fills++; } + else SET_STRING_ELT(ans->char_v, i, x[i]); + } else { SET_STRING_ELT(ans->char_v, i, x[i]); fills = 0; } } } if (verbose) @@ -117,7 +165,7 @@ void nafillString(const SEXP *x, uint_fast64_t nx, unsigned int type, SEXP fill, over columns of the input data. This includes handling different data types and applying the designated filling method to each column in parallel. */ -SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, SEXP cols) { +SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, SEXP cols, SEXP limit) { int protecti=0; const bool verbose = GetVerbose(); @@ -128,6 +176,8 @@ SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, S if (verbose) tic = omp_get_wtime(); + double limit_val = REAL(limit)[0]; + bool copy = !LOGICAL(inplace)[0]; if (!IS_TRUE_OR_FALSE(nan_is_na_arg)) error(_("'%s' must be TRUE or FALSE"), "nan_is_na"); // # nocov @@ -253,16 +303,16 @@ SEXP nafillR(SEXP obj, SEXP type, SEXP fill, SEXP nan_is_na_arg, SEXP inplace, S switch (TYPEOF(VECTOR_ELT(x, i))) { case REALSXP : { if (isInt64[i]) { - nafillInteger64(i64x[i], inx[i], itype, hasFill ? ((int64_t *)fillp[i])[0] : NA_INTEGER64, &vans[i], verbose); + nafillInteger64(i64x[i], inx[i], itype, hasFill ? ((int64_t *)fillp[i])[0] : NA_INTEGER64, &vans[i], verbose, limit_val); } else { - nafillDouble(dx[i], inx[i], itype, hasFill ? ((double *)fillp[i])[0] : NA_REAL, nan_is_na, &vans[i], verbose); + nafillDouble(dx[i], inx[i], itype, hasFill ? ((double *)fillp[i])[0] : NA_REAL, nan_is_na, &vans[i], verbose, limit_val); } } break; case LGLSXP: case INTSXP : { - nafillInteger(ix[i], inx[i], itype, hasFill ? ((int32_t *)fillp[i])[0] : NA_INTEGER, &vans[i], verbose); + nafillInteger(ix[i], inx[i], itype, hasFill ? ((int32_t *)fillp[i])[0] : NA_INTEGER, &vans[i], verbose, limit_val); } break; case STRSXP : { - nafillString(sx[i], inx[i], itype, hasFill ? ((SEXP *)fillp[i])[0] : NA_STRING, &vans[i], verbose); + nafillString(sx[i], inx[i], itype, hasFill ? ((SEXP *)fillp[i])[0] : NA_STRING, &vans[i], verbose, limit_val); } break; } }