It seems, there is a bug in exporting data on starting R session from Gretl If start a R session with the attached file opened in Gretl variables l_gdpr_usa and gdpr3 are read as they should be. gdpr_usa[1] is 99547.54  gdpr3=gdpr_usa/1000 gdpr_usa is read in R as [1] 142   1   2   5   4   6   7   8   9  15  19  21  26  27  29  31  34  37  36  40  41  43  48  [24]  46  44  50  49  52  54  58  60  64  67  70  68  72  79  81  74  82  83  85  89  91  86  88  [47]  95  94 102 103 101  99 106 112 116 118 121 122 124 131 120 115 107 104 110 113 126 130 127  [70] 132 137   3  12  17  23  20  11  10  13  14  16  18  22  25  24  28  30  33  32  35  38  39  [93]  42  45  47  51  53  55  56  57  59  61  62  66  65  63  69  71  73  76  75  78  77  80  84 [116]  87  90  92  93  96  98  97 100 105 108 109 111 114 117 119 123 125 128 129 133 134 136 140 [139] 139 135 138 141 143 The original source is  http://www.eurmacro.eu/macrodata/datamtrx.html ; the data was copy-pasted to the empty text file. Original header gdpr_usa "millions of U.S. Dollars; 1990 prices"   1870 99547.54 was edited to year gdpr_usa 1870 99547.54 and the text file was imported   If recode the data as cross-section gdpr3 is read well but gdpr_usa is read as factor instead of numeric if type as.numeric(gdpr_usa) we again get [1] 142   1   2   5 , etc.